add entrypoint to pyproject.toml

## Summary

Enables Anthropic's extended thinking feature for Claude models in
CoPilot via OpenRouter. This keeps the model's chain-of-thought
reasoning internal rather than outputting it to users.

## Problem

The CoPilot prompt was designed for a thinking agent (with
`<internal_reasoning>` tags), but extended thinking wasn't enabled on
the API side. This caused the model to output its reasoning as regular
text, leaking internal analysis to users.

## Solution

Added thinking configuration to the OpenRouter `extra_body` for
Anthropic models:
```python
extra_body["provider"] = {
    "anthropic": {
        "thinking": {
            "type": "enabled",
            "budget_tokens": config.thinking_budget_tokens,
        }
    }
}
```

## Configuration

New settings in `ChatConfig`:
| Setting | Default | Description |
|---------|---------|-------------|
| `thinking_enabled` | `True` | Enable extended thinking for Claude
models |
| `thinking_budget_tokens` | `10000` | Token budget for thinking
(1000-100000) |

## Changes

- `config.py`: Added `thinking_enabled` and `thinking_budget_tokens`
settings
- `service.py`: Added thinking config to all 3 places where `extra_body`
is built for LLM calls

## Testing

- Verify CoPilot responses no longer include internal reasoning text
- Check that Claude's extended thinking is working (should see thinking
tokens in usage)
- Confirm non-Anthropic models are unaffected

## Related

Discussion:
https://discord.com/channels/1126875755960336515/1126875756925046928/1470779843552612607

---------

Co-authored-by: Swifty <craigswift13@gmail.com>

.branchlet.json

@@ -29,8 +29,7 @@
   "postCreateCmd": [
     "cd autogpt_platform/autogpt_libs && poetry install",
     "cd autogpt_platform/backend && poetry install && poetry run prisma generate",
-    "cd autogpt_platform/frontend && pnpm install",
-    "cd docs && pip install -r requirements.txt"
+    "cd autogpt_platform/frontend && pnpm install"
   ],
   "terminalCommand": "code .",
   "deleteBranchWithWorktree": false

.github/copilot-instructions.md

@@ -160,7 +160,7 @@ pnpm storybook                      # Start component development server
 
 **Backend Entry Points:**
 
-- `backend/backend/server/server.py` - FastAPI application setup
+- `backend/backend/api/rest_api.py` - FastAPI application setup
 - `backend/backend/data/` - Database models and user management
 - `backend/blocks/` - Agent execution blocks and logic
 
@@ -219,7 +219,7 @@ Agents are built using a visual block-based system where each block performs a s
 
 ### API Development
 
-1. Update routes in `/backend/backend/server/routers/`
+1. Update routes in `/backend/backend/api/features/`
 2. Add/update Pydantic models in same directory
 3. Write tests alongside route files
 4. For `data/*.py` changes, validate user ID checks
@@ -285,7 +285,7 @@ Agents are built using a visual block-based system where each block performs a s
 
 ### Security Guidelines
 
-**Cache Protection Middleware** (`/backend/backend/server/middleware/security.py`):
+**Cache Protection Middleware** (`/backend/backend/api/middleware/security.py`):
 
 - Default: Disables caching for ALL endpoints with `Cache-Control: no-store, no-cache, must-revalidate, private`
 - Uses allow list approach for cacheable paths (static assets, health checks, public pages)

.github/workflows/classic-frontend-ci.yml

@@ -49,7 +49,7 @@ jobs:
 
       - name: Create PR ${{ env.BUILD_BRANCH }} -> ${{ github.ref_name }}
         if: github.event_name == 'push'
-        uses: peter-evans/create-pull-request@v7
+        uses: peter-evans/create-pull-request@v8
         with:
           add-paths: classic/frontend/build/web
           base: ${{ github.ref_name }}

.github/workflows/claude-ci-failure-auto-fix.yml

@@ -42,7 +42,7 @@ jobs:
 
       - name: Get CI failure details
         id: failure_details
-        uses: actions/github-script@v7
+        uses: actions/github-script@v8
         with:
           script: |
             const run = await github.rest.actions.getWorkflowRun({

.github/workflows/claude-dependabot.yml

@@ -41,7 +41,7 @@ jobs:
           python-version: "3.11"  # Use standard version matching CI
 
       - name: Set up Python dependency cache
-        uses: actions/cache@v4
+        uses: actions/cache@v5
         with:
           path: ~/.cache/pypoetry
           key: poetry-${{ runner.os }}-${{ hashFiles('autogpt_platform/backend/poetry.lock') }}
@@ -78,7 +78,7 @@ jobs:
 
       # Frontend Node.js/pnpm setup (mirrors platform-frontend-ci.yml)
       - name: Set up Node.js
-        uses: actions/setup-node@v4
+        uses: actions/setup-node@v6
         with:
           node-version: "22"
 
@@ -91,7 +91,7 @@ jobs:
           echo "PNPM_HOME=$HOME/.pnpm-store" >> $GITHUB_ENV
 
       - name: Cache frontend dependencies
-        uses: actions/cache@v4
+        uses: actions/cache@v5
         with:
           path: ~/.pnpm-store
           key: ${{ runner.os }}-pnpm-${{ hashFiles('autogpt_platform/frontend/pnpm-lock.yaml', 'autogpt_platform/frontend/package.json') }}
@@ -124,7 +124,7 @@ jobs:
       # Phase 1: Cache and load Docker images for faster setup
       - name: Set up Docker image cache
         id: docker-cache
-        uses: actions/cache@v4
+        uses: actions/cache@v5
         with:
           path: ~/docker-cache
           # Use a versioned key for cache invalidation when image list changes
@@ -309,6 +309,7 @@ jobs:
         uses: anthropics/claude-code-action@v1
         with:
           claude_code_oauth_token: ${{ secrets.CLAUDE_CODE_OAUTH_TOKEN }}
+          allowed_bots: "dependabot[bot]"
           claude_args: |
             --allowedTools "Bash(npm:*),Bash(pnpm:*),Bash(poetry:*),Bash(git:*),Edit,Replace,NotebookEditCell,mcp__github_inline_comment__create_inline_comment,Bash(gh pr comment:*), Bash(gh pr diff:*), Bash(gh pr view:*)"
           prompt: |

.github/workflows/claude.yml

@@ -57,7 +57,7 @@ jobs:
           python-version: "3.11"  # Use standard version matching CI
 
       - name: Set up Python dependency cache
-        uses: actions/cache@v4
+        uses: actions/cache@v5
         with:
           path: ~/.cache/pypoetry
           key: poetry-${{ runner.os }}-${{ hashFiles('autogpt_platform/backend/poetry.lock') }}
@@ -94,7 +94,7 @@ jobs:
 
       # Frontend Node.js/pnpm setup (mirrors platform-frontend-ci.yml)
       - name: Set up Node.js
-        uses: actions/setup-node@v4
+        uses: actions/setup-node@v6
         with:
           node-version: "22"
 
@@ -107,7 +107,7 @@ jobs:
           echo "PNPM_HOME=$HOME/.pnpm-store" >> $GITHUB_ENV
 
       - name: Cache frontend dependencies
-        uses: actions/cache@v4
+        uses: actions/cache@v5
         with:
           path: ~/.pnpm-store
           key: ${{ runner.os }}-pnpm-${{ hashFiles('autogpt_platform/frontend/pnpm-lock.yaml', 'autogpt_platform/frontend/package.json') }}
@@ -140,7 +140,7 @@ jobs:
       # Phase 1: Cache and load Docker images for faster setup
       - name: Set up Docker image cache
         id: docker-cache
-        uses: actions/cache@v4
+        uses: actions/cache@v5
         with:
           path: ~/docker-cache
           # Use a versioned key for cache invalidation when image list changes

.github/workflows/copilot-setup-steps.yml

@@ -39,7 +39,7 @@ jobs:
           python-version: "3.11"  # Use standard version matching CI
 
       - name: Set up Python dependency cache
-        uses: actions/cache@v4
+        uses: actions/cache@v5
         with:
           path: ~/.cache/pypoetry
           key: poetry-${{ runner.os }}-${{ hashFiles('autogpt_platform/backend/poetry.lock') }}
@@ -76,7 +76,7 @@ jobs:
 
       # Frontend Node.js/pnpm setup (mirrors platform-frontend-ci.yml)
       - name: Set up Node.js
-        uses: actions/setup-node@v4
+        uses: actions/setup-node@v6
         with:
           node-version: "22"
 
@@ -89,7 +89,7 @@ jobs:
           echo "PNPM_HOME=$HOME/.pnpm-store" >> $GITHUB_ENV
 
       - name: Cache frontend dependencies
-        uses: actions/cache@v4
+        uses: actions/cache@v5
         with:
           path: ~/.pnpm-store
           key: ${{ runner.os }}-pnpm-${{ hashFiles('autogpt_platform/frontend/pnpm-lock.yaml', 'autogpt_platform/frontend/package.json') }}
@@ -132,7 +132,7 @@ jobs:
       # Phase 1: Cache and load Docker images for faster setup
       - name: Set up Docker image cache
         id: docker-cache
-        uses: actions/cache@v4
+        uses: actions/cache@v5
         with:
           path: ~/docker-cache
           # Use a versioned key for cache invalidation when image list changes

.github/workflows/docs-block-sync.yml

@@ -33,7 +33,7 @@ jobs:
           python-version: "3.11"
 
       - name: Set up Python dependency cache
-        uses: actions/cache@v4
+        uses: actions/cache@v5
         with:
           path: ~/.cache/pypoetry
           key: poetry-${{ runner.os }}-${{ hashFiles('autogpt_platform/backend/poetry.lock') }}

.github/workflows/docs-claude-review.yml

@@ -33,7 +33,7 @@ jobs:
           python-version: "3.11"
 
       - name: Set up Python dependency cache
-        uses: actions/cache@v4
+        uses: actions/cache@v5
         with:
           path: ~/.cache/pypoetry
           key: poetry-${{ runner.os }}-${{ hashFiles('autogpt_platform/backend/poetry.lock') }}

.github/workflows/docs-enhance.yml

@@ -38,7 +38,7 @@ jobs:
           python-version: "3.11"
 
       - name: Set up Python dependency cache
-        uses: actions/cache@v4
+        uses: actions/cache@v5
         with:
           path: ~/.cache/pypoetry
           key: poetry-${{ runner.os }}-${{ hashFiles('autogpt_platform/backend/poetry.lock') }}

.github/workflows/platform-backend-ci.yml

@@ -88,7 +88,7 @@ jobs:
         run: echo "date=$(date +'%Y-%m-%d')" >> $GITHUB_OUTPUT
 
       - name: Set up Python dependency cache
-        uses: actions/cache@v4
+        uses: actions/cache@v5
         with:
           path: ~/.cache/pypoetry
           key: poetry-${{ runner.os }}-${{ hashFiles('autogpt_platform/backend/poetry.lock') }}

.github/workflows/platform-dev-deploy-event-dispatcher.yml

@@ -17,7 +17,7 @@ jobs:
       - name: Check comment permissions and deployment status
         id: check_status
         if: github.event_name == 'issue_comment' && github.event.issue.pull_request
-        uses: actions/github-script@v7
+        uses: actions/github-script@v8
         with:
           script: |
             const commentBody = context.payload.comment.body.trim();
@@ -55,7 +55,7 @@ jobs:
 
       - name: Post permission denied comment
         if: steps.check_status.outputs.permission_denied == 'true'
-        uses: actions/github-script@v7
+        uses: actions/github-script@v8
         with:
           script: |
             await github.rest.issues.createComment({
@@ -68,7 +68,7 @@ jobs:
       - name: Get PR details for deployment
         id: pr_details
         if: steps.check_status.outputs.should_deploy == 'true' || steps.check_status.outputs.should_undeploy == 'true'
-        uses: actions/github-script@v7
+        uses: actions/github-script@v8
         with:
           script: |
             const pr = await github.rest.pulls.get({
@@ -98,7 +98,7 @@ jobs:
 
       - name: Post deploy success comment
         if: steps.check_status.outputs.should_deploy == 'true'
-        uses: actions/github-script@v7
+        uses: actions/github-script@v8
         with:
           script: |
             await github.rest.issues.createComment({
@@ -126,7 +126,7 @@ jobs:
 
       - name: Post undeploy success comment
         if: steps.check_status.outputs.should_undeploy == 'true'
-        uses: actions/github-script@v7
+        uses: actions/github-script@v8
         with:
           script: |
             await github.rest.issues.createComment({
@@ -139,7 +139,7 @@ jobs:
       - name: Check deployment status on PR close
         id: check_pr_close
         if: github.event_name == 'pull_request' && github.event.action == 'closed'
-        uses: actions/github-script@v7
+        uses: actions/github-script@v8
         with:
           script: |
             const comments = await github.rest.issues.listComments({
@@ -187,7 +187,7 @@ jobs:
           github.event_name == 'pull_request' &&
           github.event.action == 'closed' &&
           steps.check_pr_close.outputs.should_undeploy == 'true'
-        uses: actions/github-script@v7
+        uses: actions/github-script@v8
         with:
           script: |
             await github.rest.issues.createComment({

.github/workflows/platform-frontend-ci.yml

@@ -27,13 +27,22 @@ jobs:
     runs-on: ubuntu-latest
     outputs:
       cache-key: ${{ steps.cache-key.outputs.key }}
+      components-changed: ${{ steps.filter.outputs.components }}
 
     steps:
       - name: Checkout repository
         uses: actions/checkout@v4
 
+      - name: Check for component changes
+        uses: dorny/paths-filter@v3
+        id: filter
+        with:
+          filters: |
+            components:
+              - 'autogpt_platform/frontend/src/components/**'
+
       - name: Set up Node.js
-        uses: actions/setup-node@v4
+        uses: actions/setup-node@v6
         with:
           node-version: "22.18.0"
 
@@ -45,7 +54,7 @@ jobs:
         run: echo "key=${{ runner.os }}-pnpm-${{ hashFiles('autogpt_platform/frontend/pnpm-lock.yaml', 'autogpt_platform/frontend/package.json') }}" >> $GITHUB_OUTPUT
 
       - name: Cache dependencies
-        uses: actions/cache@v4
+        uses: actions/cache@v5
         with:
           path: ~/.pnpm-store
           key: ${{ steps.cache-key.outputs.key }}
@@ -65,7 +74,7 @@ jobs:
         uses: actions/checkout@v4
 
       - name: Set up Node.js
-        uses: actions/setup-node@v4
+        uses: actions/setup-node@v6
         with:
           node-version: "22.18.0"
 
@@ -73,7 +82,7 @@ jobs:
         run: corepack enable
 
       - name: Restore dependencies cache
-        uses: actions/cache@v4
+        uses: actions/cache@v5
         with:
           path: ~/.pnpm-store
           key: ${{ needs.setup.outputs.cache-key }}
@@ -90,8 +99,11 @@ jobs:
   chromatic:
     runs-on: ubuntu-latest
     needs: setup
-    # Only run on dev branch pushes or PRs targeting dev
-    if: github.ref == 'refs/heads/dev' || github.base_ref == 'dev'
+    # Disabled: to re-enable, remove 'false &&' from the condition below
+    if: >-
+      false
+      && (github.ref == 'refs/heads/dev' || github.base_ref == 'dev')
+      && needs.setup.outputs.components-changed == 'true'
 
     steps:
       - name: Checkout repository
@@ -100,7 +112,7 @@ jobs:
           fetch-depth: 0
 
       - name: Set up Node.js
-        uses: actions/setup-node@v4
+        uses: actions/setup-node@v6
         with:
           node-version: "22.18.0"
 
@@ -108,7 +120,7 @@ jobs:
         run: corepack enable
 
       - name: Restore dependencies cache
-        uses: actions/cache@v4
+        uses: actions/cache@v5
         with:
           path: ~/.pnpm-store
           key: ${{ needs.setup.outputs.cache-key }}
@@ -141,7 +153,7 @@ jobs:
           submodules: recursive
 
       - name: Set up Node.js
-        uses: actions/setup-node@v4
+        uses: actions/setup-node@v6
         with:
           node-version: "22.18.0"
 
@@ -164,7 +176,7 @@ jobs:
         uses: docker/setup-buildx-action@v3
 
       - name: Cache Docker layers
-        uses: actions/cache@v4
+        uses: actions/cache@v5
         with:
           path: /tmp/.buildx-cache
           key: ${{ runner.os }}-buildx-frontend-test-${{ hashFiles('autogpt_platform/docker-compose.yml', 'autogpt_platform/backend/Dockerfile', 'autogpt_platform/backend/pyproject.toml', 'autogpt_platform/backend/poetry.lock') }}
@@ -219,7 +231,7 @@ jobs:
           fi
 
       - name: Restore dependencies cache
-        uses: actions/cache@v4
+        uses: actions/cache@v5
         with:
           path: ~/.pnpm-store
           key: ${{ needs.setup.outputs.cache-key }}
@@ -270,7 +282,7 @@ jobs:
           submodules: recursive
 
       - name: Set up Node.js
-        uses: actions/setup-node@v4
+        uses: actions/setup-node@v6
         with:
           node-version: "22.18.0"
 
@@ -278,7 +290,7 @@ jobs:
         run: corepack enable
 
       - name: Restore dependencies cache
-        uses: actions/cache@v4
+        uses: actions/cache@v5
         with:
           path: ~/.pnpm-store
           key: ${{ needs.setup.outputs.cache-key }}

.github/workflows/platform-fullstack-ci.yml

@@ -32,7 +32,7 @@ jobs:
         uses: actions/checkout@v4
 
       - name: Set up Node.js
-        uses: actions/setup-node@v4
+        uses: actions/setup-node@v6
         with:
           node-version: "22.18.0"
 
@@ -44,7 +44,7 @@ jobs:
         run: echo "key=${{ runner.os }}-pnpm-${{ hashFiles('autogpt_platform/frontend/pnpm-lock.yaml', 'autogpt_platform/frontend/package.json') }}" >> $GITHUB_OUTPUT
 
       - name: Cache dependencies
-        uses: actions/cache@v4
+        uses: actions/cache@v5
         with:
           path: ~/.pnpm-store
           key: ${{ steps.cache-key.outputs.key }}
@@ -56,7 +56,7 @@ jobs:
         run: pnpm install --frozen-lockfile
 
   types:
-    runs-on: ubuntu-latest
+    runs-on: big-boi
     needs: setup
     strategy:
       fail-fast: false
@@ -68,7 +68,7 @@ jobs:
           submodules: recursive
 
       - name: Set up Node.js
-        uses: actions/setup-node@v4
+        uses: actions/setup-node@v6
         with:
           node-version: "22.18.0"
 
@@ -85,10 +85,10 @@ jobs:
 
       - name: Run docker compose
         run: |
-          docker compose -f ../docker-compose.yml --profile local --profile deps_backend up -d
+          docker compose -f ../docker-compose.yml --profile local up -d deps_backend
 
       - name: Restore dependencies cache
-        uses: actions/cache@v4
+        uses: actions/cache@v5
         with:
           path: ~/.pnpm-store
           key: ${{ needs.setup.outputs.cache-key }}

.gitignore

@@ -178,4 +178,6 @@ autogpt_platform/backend/settings.py
 *.ign.*
 .test-contents
 .claude/settings.local.json
+CLAUDE.local.md
 /autogpt_platform/backend/logs
+.next

AGENTS.md

@@ -16,7 +16,6 @@ See `docs/content/platform/getting-started.md` for setup instructions.
 - Format Python code with `poetry run format`.
 - Format frontend code using `pnpm format`.
 
-
 ## Frontend guidelines:
 
 See `/frontend/CONTRIBUTING.md` for complete patterns. Quick reference:
@@ -33,14 +32,17 @@ See `/frontend/CONTRIBUTING.md` for complete patterns. Quick reference:
 4. **Styling**: Tailwind CSS only, use design tokens, Phosphor Icons only
 5. **Testing**: Add Storybook stories for new components, Playwright for E2E
 6. **Code conventions**: Function declarations (not arrow functions) for components/handlers
+
 - Component props should be `interface Props { ... }` (not exported) unless the interface needs to be used outside the component
 - Separate render logic from business logic (component.tsx + useComponent.ts + helpers.ts)
 - Colocate state when possible and avoid creating large components, use sub-components ( local `/components` folder next to the parent component ) when sensible
 - Avoid large hooks, abstract logic into `helpers.ts` files when sensible
 - Use function declarations for components, arrow functions only for callbacks
 - No barrel files or `index.ts` re-exports
-- Do not use `useCallback` or `useMemo` unless strictly needed
 - Avoid comments at all times unless the code is very complex
+- Do not use `useCallback` or `useMemo` unless asked to optimise a given function
+- Do not type hook returns, let Typescript infer as much as possible
+- Never type with `any`, if not types available use `unknown`
 
 ## Testing
 
@@ -49,22 +51,8 @@ See `/frontend/CONTRIBUTING.md` for complete patterns. Quick reference:
 
 Always run the relevant linters and tests before committing.
 Use conventional commit messages for all commits (e.g. `feat(backend): add API`).
-  Types:
-    - feat
-    - fix
-    - refactor
-    - ci
-    - dx (developer experience)
-  Scopes:
-    - platform
-      - platform/library
-      - platform/marketplace
-      - backend
-        - backend/executor
-      - frontend
-        - frontend/library
-        - frontend/marketplace
-      - blocks
+Types: - feat - fix - refactor - ci - dx (developer experience)
+Scopes: - platform - platform/library - platform/marketplace - backend - backend/executor - frontend - frontend/library - frontend/marketplace - blocks
 
 ## Pull requests
 

README.md

@@ -54,7 +54,7 @@ Before proceeding with the installation, ensure your system meets the following
 ### Updated Setup Instructions:
 We've moved to a fully maintained and regularly updated documentation site.
 
-👉 [Follow the official self-hosting guide here](https://docs.agpt.co/platform/getting-started/)
+👉 [Follow the official self-hosting guide here](https://agpt.co/docs/platform/getting-started/getting-started)
 
 
 This tutorial assumes you have Docker, VSCode, git and npm installed.

autogpt_platform/CLAUDE.md

@@ -6,152 +6,30 @@ This file provides guidance to Claude Code (claude.ai/code) when working with co
 
 AutoGPT Platform is a monorepo containing:
 
-- **Backend** (`/backend`): Python FastAPI server with async support
-- **Frontend** (`/frontend`): Next.js React application
-- **Shared Libraries** (`/autogpt_libs`): Common Python utilities
+- **Backend** (`backend`): Python FastAPI server with async support
+- **Frontend** (`frontend`): Next.js React application
+- **Shared Libraries** (`autogpt_libs`): Common Python utilities
 
-## Essential Commands
+## Component Documentation
 
-### Backend Development
+- **Backend**: See @backend/CLAUDE.md for backend-specific commands, architecture, and development tasks
+- **Frontend**: See @frontend/CLAUDE.md for frontend-specific commands, architecture, and development patterns
 
-```bash
-# Install dependencies
-cd backend && poetry install
-
-# Run database migrations
-poetry run prisma migrate dev
-
-# Start all services (database, redis, rabbitmq, clamav)
-docker compose up -d
-
-# Run the backend server
-poetry run serve
-
-# Run tests
-poetry run test
-
-# Run specific test
-poetry run pytest path/to/test_file.py::test_function_name
-
-# Run block tests (tests that validate all blocks work correctly)
-poetry run pytest backend/blocks/test/test_block.py -xvs
-
-# Run tests for a specific block (e.g., GetCurrentTimeBlock)
-poetry run pytest 'backend/blocks/test/test_block.py::test_available_blocks[GetCurrentTimeBlock]' -xvs
-
-# Lint and format
-# prefer format if you want to just "fix" it and only get the errors that can't be autofixed
-poetry run format  # Black + isort
-poetry run lint    # ruff
-```
-
-More details can be found in TESTING.md
-
-#### Creating/Updating Snapshots
-
-When you first write a test or when the expected output changes:
-
-```bash
-poetry run pytest path/to/test.py --snapshot-update
-```
-
-⚠️ **Important**: Always review snapshot changes before committing! Use `git diff` to verify the changes are expected.
-
-### Frontend Development
-
-```bash
-# Install dependencies
-cd frontend && pnpm i
-
-# Generate API client from OpenAPI spec
-pnpm generate:api
-
-# Start development server
-pnpm dev
-
-# Run E2E tests
-pnpm test
-
-# Run Storybook for component development
-pnpm storybook
-
-# Build production
-pnpm build
-
-# Format and lint
-pnpm format
-
-# Type checking
-pnpm types
-```
-
-**📖 Complete Guide**: See `/frontend/CONTRIBUTING.md` and `/frontend/.cursorrules` for comprehensive frontend patterns.
-
-**Key Frontend Conventions:**
-
-- Separate render logic from data/behavior in components
-- Use generated API hooks from `@/app/api/__generated__/endpoints/`
-- Use function declarations (not arrow functions) for components/handlers
-- Use design system components from `src/components/` (atoms, molecules, organisms)
-- Only use Phosphor Icons
-- Never use `src/components/__legacy__/*` or deprecated `BackendAPI`
-
-## Architecture Overview
-
-### Backend Architecture
-
-- **API Layer**: FastAPI with REST and WebSocket endpoints
-- **Database**: PostgreSQL with Prisma ORM, includes pgvector for embeddings
-- **Queue System**: RabbitMQ for async task processing
-- **Execution Engine**: Separate executor service processes agent workflows
-- **Authentication**: JWT-based with Supabase integration
-- **Security**: Cache protection middleware prevents sensitive data caching in browsers/proxies
-
-### Frontend Architecture
-
-- **Framework**: Next.js 15 App Router (client-first approach)
-- **Data Fetching**: Type-safe generated API hooks via Orval + React Query
-- **State Management**: React Query for server state, co-located UI state in components/hooks
-- **Component Structure**: Separate render logic (`.tsx`) from business logic (`use*.ts` hooks)
-- **Workflow Builder**: Visual graph editor using @xyflow/react
-- **UI Components**: shadcn/ui (Radix UI primitives) with Tailwind CSS styling
-- **Icons**: Phosphor Icons only
-- **Feature Flags**: LaunchDarkly integration
-- **Error Handling**: ErrorCard for render errors, toast for mutations, Sentry for exceptions
-- **Testing**: Playwright for E2E, Storybook for component development
-
-### Key Concepts
+## Key Concepts
 
 1. **Agent Graphs**: Workflow definitions stored as JSON, executed by the backend
-2. **Blocks**: Reusable components in `/backend/blocks/` that perform specific tasks
+2. **Blocks**: Reusable components in `backend/backend/blocks/` that perform specific tasks
 3. **Integrations**: OAuth and API connections stored per user
 4. **Store**: Marketplace for sharing agent templates
 5. **Virus Scanning**: ClamAV integration for file upload security
 
-### Testing Approach
-
-- Backend uses pytest with snapshot testing for API responses
-- Test files are colocated with source files (`*_test.py`)
-- Frontend uses Playwright for E2E tests
-- Component testing via Storybook
-
-### Database Schema
-
-Key models (defined in `/backend/schema.prisma`):
-
-- `User`: Authentication and profile data
-- `AgentGraph`: Workflow definitions with version control
-- `AgentGraphExecution`: Execution history and results
-- `AgentNode`: Individual nodes in a workflow
-- `StoreListing`: Marketplace listings for sharing agents
-
 ### Environment Configuration
 
 #### Configuration Files
 
-- **Backend**: `/backend/.env.default` (defaults) → `/backend/.env` (user overrides)
-- **Frontend**: `/frontend/.env.default` (defaults) → `/frontend/.env` (user overrides)
-- **Platform**: `/.env.default` (Supabase/shared defaults) → `/.env` (user overrides)
+- **Backend**: `backend/.env.default` (defaults) → `backend/.env` (user overrides)
+- **Frontend**: `frontend/.env.default` (defaults) → `frontend/.env` (user overrides)
+- **Platform**: `.env.default` (Supabase/shared defaults) → `.env` (user overrides)
 
 #### Docker Environment Loading Order
 
@@ -167,83 +45,12 @@ Key models (defined in `/backend/schema.prisma`):
 - Backend/Frontend services use YAML anchors for consistent configuration
 - Supabase services (`db/docker/docker-compose.yml`) follow the same pattern
 
-### Common Development Tasks
-
-**Adding a new block:**
-
-Follow the comprehensive [Block SDK Guide](../../../docs/content/platform/block-sdk-guide.md) which covers:
-
-- Provider configuration with `ProviderBuilder`
-- Block schema definition
-- Authentication (API keys, OAuth, webhooks)
-- Testing and validation
-- File organization
-
-Quick steps:
-
-1. Create new file in `/backend/backend/blocks/`
-2. Configure provider using `ProviderBuilder` in `_config.py`
-3. Inherit from `Block` base class
-4. Define input/output schemas using `BlockSchema`
-5. Implement async `run` method
-6. Generate unique block ID using `uuid.uuid4()`
-7. Test with `poetry run pytest backend/blocks/test/test_block.py`
-
-Note: when making many new blocks analyze the interfaces for each of these blocks and picture if they would go well together in a graph based editor or would they struggle to connect productively?
-ex: do the inputs and outputs tie well together?
-
-If you get any pushback or hit complex block conditions check the new_blocks guide in the docs.
-
-**Modifying the API:**
-
-1. Update route in `/backend/backend/server/routers/`
-2. Add/update Pydantic models in same directory
-3. Write tests alongside the route file
-4. Run `poetry run test` to verify
-
-### Frontend guidelines:
-
-See `/frontend/CONTRIBUTING.md` for complete patterns. Quick reference:
-
-1. **Pages**: Create in `src/app/(platform)/feature-name/page.tsx`
-   - Add `usePageName.ts` hook for logic
-   - Put sub-components in local `components/` folder
-2. **Components**: Structure as `ComponentName/ComponentName.tsx` + `useComponentName.ts` + `helpers.ts`
-   - Use design system components from `src/components/` (atoms, molecules, organisms)
-   - Never use `src/components/__legacy__/*`
-3. **Data fetching**: Use generated API hooks from `@/app/api/__generated__/endpoints/`
-   - Regenerate with `pnpm generate:api`
-   - Pattern: `use{Method}{Version}{OperationName}`
-4. **Styling**: Tailwind CSS only, use design tokens, Phosphor Icons only
-5. **Testing**: Add Storybook stories for new components, Playwright for E2E
-6. **Code conventions**: Function declarations (not arrow functions) for components/handlers
-- Component props should be `interface Props { ... }` (not exported) unless the interface needs to be used outside the component
-- Separate render logic from business logic (component.tsx + useComponent.ts + helpers.ts)
-- Colocate state when possible and avoid creating large components, use sub-components ( local `/components` folder next to the parent component ) when sensible
-- Avoid large hooks, abstract logic into `helpers.ts` files when sensible
-- Use function declarations for components, arrow functions only for callbacks
-- No barrel files or `index.ts` re-exports
-- Do not use `useCallback` or `useMemo` unless strictly needed
-- Avoid comments at all times unless the code is very complex
-
-### Security Implementation
-
-**Cache Protection Middleware:**
-
-- Located in `/backend/backend/server/middleware/security.py`
-- Default behavior: Disables caching for ALL endpoints with `Cache-Control: no-store, no-cache, must-revalidate, private`
-- Uses an allow list approach - only explicitly permitted paths can be cached
-- Cacheable paths include: static assets (`/static/*`, `/_next/static/*`), health checks, public store pages, documentation
-- Prevents sensitive data (auth tokens, API keys, user data) from being cached by browsers/proxies
-- To allow caching for a new endpoint, add it to `CACHEABLE_PATHS` in the middleware
-- Applied to both main API server and external API applications
-
 ### Creating Pull Requests
 
-- Create the PR aginst the `dev` branch of the repository.
-- Ensure the branch name is descriptive (e.g., `feature/add-new-block`)/
-- Use conventional commit messages (see below)/
-- Fill out the .github/PULL_REQUEST_TEMPLATE.md template as the PR description/
+- Create the PR against the `dev` branch of the repository.
+- Ensure the branch name is descriptive (e.g., `feature/add-new-block`)
+- Use conventional commit messages (see below)
+- Fill out the .github/PULL_REQUEST_TEMPLATE.md template as the PR description
 - Run the github pre-commit hooks to ensure code quality.
 
 ### Reviewing/Revising Pull Requests

autogpt_platform/autogpt_libs/pyproject.toml

@@ -9,25 +9,25 @@ packages = [{ include = "autogpt_libs" }]
 [tool.poetry.dependencies]
 python = ">=3.10,<4.0"
 colorama = "^0.4.6"
-cryptography = "^45.0"
+cryptography = "^46.0"
 expiringdict = "^1.2.2"
-fastapi = "^0.116.1"
-google-cloud-logging = "^3.12.1"
-launchdarkly-server-sdk = "^9.12.0"
-pydantic = "^2.11.7"
-pydantic-settings = "^2.10.1"
-pyjwt = { version = "^2.10.1", extras = ["crypto"] }
+fastapi = "^0.128.0"
+google-cloud-logging = "^3.13.0"
+launchdarkly-server-sdk = "^9.14.1"
+pydantic = "^2.12.5"
+pydantic-settings = "^2.12.0"
+pyjwt = { version = "^2.11.0", extras = ["crypto"] }
 redis = "^6.2.0"
-supabase = "^2.16.0"
-uvicorn = "^0.35.0"
+supabase = "^2.27.2"
+uvicorn = "^0.40.0"
 
 [tool.poetry.group.dev.dependencies]
-pyright = "^1.1.404"
+pyright = "^1.1.408"
 pytest = "^8.4.1"
-pytest-asyncio = "^1.1.0"
-pytest-mock = "^3.14.1"
-pytest-cov = "^6.2.1"
-ruff = "^0.12.11"
+pytest-asyncio = "^1.3.0"
+pytest-mock = "^3.15.1"
+pytest-cov = "^7.0.0"
+ruff = "^0.15.0"
 
 [build-system]
 requires = ["poetry-core"]

autogpt_platform/backend/.env.default

@@ -152,6 +152,7 @@ REPLICATE_API_KEY=
 REVID_API_KEY=
 SCREENSHOTONE_API_KEY=
 UNREAL_SPEECH_API_KEY=
+ELEVENLABS_API_KEY=
 
 # Data & Search Services
 E2B_API_KEY=
@@ -178,5 +179,10 @@ AYRSHARE_JWT_KEY=
 SMARTLEAD_API_KEY=
 ZEROBOUNCE_API_KEY=
 
+# PostHog Analytics
+# Get API key from https://posthog.com - Project Settings > Project API Key
+POSTHOG_API_KEY=
+POSTHOG_HOST=https://eu.i.posthog.com
+
 # Other Services
 AUTOMOD_API_KEY=

autogpt_platform/backend/.gitignore

@@ -19,3 +19,6 @@ load-tests/*.json
 load-tests/*.log
 load-tests/node_modules/*
 migrations/*/rollback*.sql
+
+# Workspace files
+workspaces/

autogpt_platform/backend/CLAUDE.md

@@ -0,0 +1,170 @@
+# CLAUDE.md - Backend
+
+This file provides guidance to Claude Code when working with the backend.
+
+## Essential Commands
+
+To run something with Python package dependencies you MUST use `poetry run ...`.
+
+```bash
+# Install dependencies
+poetry install
+
+# Run database migrations
+poetry run prisma migrate dev
+
+# Start all services (database, redis, rabbitmq, clamav)
+docker compose up -d
+
+# Run the backend as a whole
+poetry run app
+
+# Run tests
+poetry run test
+
+# Run specific test
+poetry run pytest path/to/test_file.py::test_function_name
+
+# Run block tests (tests that validate all blocks work correctly)
+poetry run pytest backend/blocks/test/test_block.py -xvs
+
+# Run tests for a specific block (e.g., GetCurrentTimeBlock)
+poetry run pytest 'backend/blocks/test/test_block.py::test_available_blocks[GetCurrentTimeBlock]' -xvs
+
+# Lint and format
+# prefer format if you want to just "fix" it and only get the errors that can't be autofixed
+poetry run format  # Black + isort
+poetry run lint    # ruff
+```
+
+More details can be found in @TESTING.md
+
+### Creating/Updating Snapshots
+
+When you first write a test or when the expected output changes:
+
+```bash
+poetry run pytest path/to/test.py --snapshot-update
+```
+
+⚠️ **Important**: Always review snapshot changes before committing! Use `git diff` to verify the changes are expected.
+
+## Architecture
+
+- **API Layer**: FastAPI with REST and WebSocket endpoints
+- **Database**: PostgreSQL with Prisma ORM, includes pgvector for embeddings
+- **Queue System**: RabbitMQ for async task processing
+- **Execution Engine**: Separate executor service processes agent workflows
+- **Authentication**: JWT-based with Supabase integration
+- **Security**: Cache protection middleware prevents sensitive data caching in browsers/proxies
+
+## Testing Approach
+
+- Uses pytest with snapshot testing for API responses
+- Test files are colocated with source files (`*_test.py`)
+
+## Database Schema
+
+Key models (defined in `schema.prisma`):
+
+- `User`: Authentication and profile data
+- `AgentGraph`: Workflow definitions with version control
+- `AgentGraphExecution`: Execution history and results
+- `AgentNode`: Individual nodes in a workflow
+- `StoreListing`: Marketplace listings for sharing agents
+
+## Environment Configuration
+
+- **Backend**: `.env.default` (defaults) → `.env` (user overrides)
+
+## Common Development Tasks
+
+### Adding a new block
+
+Follow the comprehensive [Block SDK Guide](@../../docs/content/platform/block-sdk-guide.md) which covers:
+
+- Provider configuration with `ProviderBuilder`
+- Block schema definition
+- Authentication (API keys, OAuth, webhooks)
+- Testing and validation
+- File organization
+
+Quick steps:
+
+1. Create new file in `backend/blocks/`
+2. Configure provider using `ProviderBuilder` in `_config.py`
+3. Inherit from `Block` base class
+4. Define input/output schemas using `BlockSchema`
+5. Implement async `run` method
+6. Generate unique block ID using `uuid.uuid4()`
+7. Test with `poetry run pytest backend/blocks/test/test_block.py`
+
+Note: when making many new blocks analyze the interfaces for each of these blocks and picture if they would go well together in a graph-based editor or would they struggle to connect productively?
+ex: do the inputs and outputs tie well together?
+
+If you get any pushback or hit complex block conditions check the new_blocks guide in the docs.
+
+#### Handling files in blocks with `store_media_file()`
+
+When blocks need to work with files (images, videos, documents), use `store_media_file()` from `backend.util.file`. The `return_format` parameter determines what you get back:
+
+| Format | Use When | Returns |
+|--------|----------|---------|
+| `"for_local_processing"` | Processing with local tools (ffmpeg, MoviePy, PIL) | Local file path (e.g., `"image.png"`) |
+| `"for_external_api"` | Sending content to external APIs (Replicate, OpenAI) | Data URI (e.g., `"data:image/png;base64,..."`) |
+| `"for_block_output"` | Returning output from your block | Smart: `workspace://` in CoPilot, data URI in graphs |
+
+**Examples:**
+
+```python
+# INPUT: Need to process file locally with ffmpeg
+local_path = await store_media_file(
+    file=input_data.video,
+    execution_context=execution_context,
+    return_format="for_local_processing",
+)
+# local_path = "video.mp4" - use with Path/ffmpeg/etc
+
+# INPUT: Need to send to external API like Replicate
+image_b64 = await store_media_file(
+    file=input_data.image,
+    execution_context=execution_context,
+    return_format="for_external_api",
+)
+# image_b64 = "data:image/png;base64,iVBORw0..." - send to API
+
+# OUTPUT: Returning result from block
+result_url = await store_media_file(
+    file=generated_image_url,
+    execution_context=execution_context,
+    return_format="for_block_output",
+)
+yield "image_url", result_url
+# In CoPilot: result_url = "workspace://abc123"
+# In graphs:  result_url = "data:image/png;base64,..."
+```
+
+**Key points:**
+
+- `for_block_output` is the ONLY format that auto-adapts to execution context
+- Always use `for_block_output` for block outputs unless you have a specific reason not to
+- Never hardcode workspace checks - let `for_block_output` handle it
+
+### Modifying the API
+
+1. Update route in `backend/api/features/`
+2. Add/update Pydantic models in same directory
+3. Write tests alongside the route file
+4. Run `poetry run test` to verify
+
+## Security Implementation
+
+### Cache Protection Middleware
+
+- Located in `backend/api/middleware/security.py`
+- Default behavior: Disables caching for ALL endpoints with `Cache-Control: no-store, no-cache, must-revalidate, private`
+- Uses an allow list approach - only explicitly permitted paths can be cached
+- Cacheable paths include: static assets (`static/*`, `_next/static/*`), health checks, public store pages, documentation
+- Prevents sensitive data (auth tokens, API keys, user data) from being cached by browsers/proxies
+- To allow caching for a new endpoint, add it to `CACHEABLE_PATHS` in the middleware
+- Applied to both main API server and external API applications

autogpt_platform/backend/Dockerfile

@@ -62,10 +62,12 @@ ENV POETRY_HOME=/opt/poetry \
     DEBIAN_FRONTEND=noninteractive
 ENV PATH=/opt/poetry/bin:$PATH
 
-# Install Python without upgrading system-managed packages
+# Install Python, FFmpeg, and ImageMagick (required for video processing blocks)
 RUN apt-get update && apt-get install -y \
     python3.13 \
     python3-pip \
+    ffmpeg \
+    imagemagick \
     && rm -rf /var/lib/apt/lists/*
 
 # Copy only necessary files from builder

autogpt_platform/backend/TESTING.md

@@ -138,7 +138,7 @@ If the test doesn't need the `user_id` specifically, mocking is not necessary as
 
 #### Using Global Auth Fixtures
 
-Two global auth fixtures are provided by `backend/server/conftest.py`:
+Two global auth fixtures are provided by `backend/api/conftest.py`:
 
 - `mock_jwt_user` - Regular user with `test_user_id` ("test-user-id")
 - `mock_jwt_admin` - Admin user with `admin_user_id` ("admin-user-id")

autogpt_platform/backend/backend/api/external/v1/routes.py

@@ -86,6 +86,8 @@ async def execute_graph_block(
     obj = backend.data.block.get_block(block_id)
     if not obj:
         raise HTTPException(status_code=404, detail=f"Block #{block_id} not found.")
+    if obj.disabled:
+        raise HTTPException(status_code=403, detail=f"Block #{block_id} is disabled.")
 
     output = defaultdict(list)
     async for name, data in obj.execute(data):

autogpt_platform/backend/backend/api/external/v1/tools.py

@@ -15,9 +15,9 @@ from prisma.enums import APIKeyPermission
 from pydantic import BaseModel, Field
 
 from backend.api.external.middleware import require_permission
-from backend.api.features.chat.model import ChatSession
-from backend.api.features.chat.tools import find_agent_tool, run_agent_tool
-from backend.api.features.chat.tools.models import ToolResponseBase
+from backend.copilot.model import ChatSession
+from backend.copilot.tools import find_agent_tool, run_agent_tool
+from backend.copilot.tools.models import ToolResponseBase
 from backend.data.auth.base import APIAuthorizationInfo
 
 logger = logging.getLogger(__name__)

autogpt_platform/backend/backend/api/features/builder/routes.py

@@ -17,7 +17,7 @@ router = fastapi.APIRouter(
 )
 
 
-# Taken from backend/server/v2/store/db.py
+# Taken from backend/api/features/store/db.py
 def sanitize_query(query: str | None) -> str | None:
     if query is None:
         return query

autogpt_platform/backend/backend/api/features/chat/routes.py

@@ -1,20 +1,54 @@
 """Chat API routes for chat session management and streaming via SSE."""
 
 import logging
+import uuid as uuid_module
 from collections.abc import AsyncGenerator
 from typing import Annotated
 
 from autogpt_libs import auth
-from fastapi import APIRouter, Depends, Query, Security
+from fastapi import APIRouter, Depends, Header, HTTPException, Query, Response, Security
 from fastapi.responses import StreamingResponse
 from pydantic import BaseModel
 
+from backend.copilot import service as chat_service
+from backend.copilot import stream_registry
+from backend.copilot.completion_handler import (
+    process_operation_failure,
+    process_operation_success,
+)
+from backend.copilot.config import ChatConfig
+from backend.copilot.executor.utils import enqueue_copilot_task
+from backend.copilot.model import (
+    ChatSession,
+    create_chat_session,
+    get_chat_session,
+    get_user_sessions,
+)
+from backend.copilot.response_model import StreamFinish, StreamHeartbeat
+from backend.copilot.tools.models import (
+    AgentDetailsResponse,
+    AgentOutputResponse,
+    AgentPreviewResponse,
+    AgentSavedResponse,
+    AgentsFoundResponse,
+    BlockListResponse,
+    BlockOutputResponse,
+    ClarificationNeededResponse,
+    DocPageResponse,
+    DocSearchResultsResponse,
+    ErrorResponse,
+    ExecutionStartedResponse,
+    InputValidationErrorResponse,
+    NeedLoginResponse,
+    NoResultsResponse,
+    OperationInProgressResponse,
+    OperationPendingResponse,
+    OperationStartedResponse,
+    SetupRequirementsResponse,
+    UnderstandingUpdatedResponse,
+)
 from backend.util.exceptions import NotFoundError
 
-from . import service as chat_service
-from .config import ChatConfig
-from .model import ChatSession, create_chat_session, get_chat_session, get_user_sessions
-
 config = ChatConfig()
 
 
@@ -55,6 +89,15 @@ class CreateSessionResponse(BaseModel):
     user_id: str | None
 
 
+class ActiveStreamInfo(BaseModel):
+    """Information about an active stream for reconnection."""
+
+    task_id: str
+    last_message_id: str  # Redis Stream message ID for resumption
+    operation_id: str  # Operation ID for completion tracking
+    tool_name: str  # Name of the tool being executed
+
+
 class SessionDetailResponse(BaseModel):
     """Response model providing complete details for a chat session, including messages."""
 
@@ -63,6 +106,7 @@ class SessionDetailResponse(BaseModel):
     updated_at: str
     user_id: str | None
     messages: list[dict]
+    active_stream: ActiveStreamInfo | None = None  # Present if stream is still active
 
 
 class SessionSummaryResponse(BaseModel):
@@ -81,6 +125,14 @@ class ListSessionsResponse(BaseModel):
     total: int
 
 
+class OperationCompleteRequest(BaseModel):
+    """Request model for external completion webhook."""
+
+    success: bool
+    result: dict | str | None = None
+    error: str | None = None
+
+
 # ========== Routes ==========
 
 
@@ -166,13 +218,14 @@ async def get_session(
     Retrieve the details of a specific chat session.
 
     Looks up a chat session by ID for the given user (if authenticated) and returns all session data including messages.
+    If there's an active stream for this session, returns the task_id for reconnection.
 
     Args:
         session_id: The unique identifier for the desired chat session.
         user_id: The optional authenticated user ID, or None for anonymous access.
 
     Returns:
-        SessionDetailResponse: Details for the requested session, or None if not found.
+        SessionDetailResponse: Details for the requested session, including active_stream info if applicable.
 
     """
     session = await get_chat_session(session_id, user_id)
@@ -180,11 +233,28 @@ async def get_session(
         raise NotFoundError(f"Session {session_id} not found.")
 
     messages = [message.model_dump() for message in session.messages]
-    logger.info(
-        f"Returning session {session_id}: "
-        f"message_count={len(messages)}, "
-        f"roles={[m.get('role') for m in messages]}"
+
+    # Check if there's an active stream for this session
+    active_stream_info = None
+    active_task, last_message_id = await stream_registry.get_active_task_for_session(
+        session_id, user_id
     )
+    if active_task:
+        # Filter out the in-progress assistant message from the session response.
+        # The client will receive the complete assistant response through the SSE
+        # stream replay instead, preventing duplicate content.
+        if messages and messages[-1].get("role") == "assistant":
+            messages = messages[:-1]
+
+        # Use "0-0" as last_message_id to replay the stream from the beginning.
+        # Since we filtered out the cached assistant message, the client needs
+        # the full stream to reconstruct the response.
+        active_stream_info = ActiveStreamInfo(
+            task_id=active_task.task_id,
+            last_message_id="0-0",
+            operation_id=active_task.operation_id,
+            tool_name=active_task.tool_name,
+        )
 
     return SessionDetailResponse(
         id=session.session_id,
@@ -192,6 +262,7 @@ async def get_session(
         updated_at=session.updated_at.isoformat(),
         user_id=session.user_id or None,
         messages=messages,
+        active_stream=active_stream_info,
     )
 
 
@@ -211,49 +282,202 @@ async def stream_chat_post(
       - Tool call UI elements (if invoked)
       - Tool execution results
 
+    The AI generation runs in a background task that continues even if the client disconnects.
+    All chunks are written to Redis for reconnection support. If the client disconnects,
+    they can reconnect using GET /tasks/{task_id}/stream to resume from where they left off.
+
     Args:
         session_id: The chat session identifier to associate with the streamed messages.
         request: Request body containing message, is_user_message, and optional context.
         user_id: Optional authenticated user ID.
     Returns:
-        StreamingResponse: SSE-formatted response chunks.
+        StreamingResponse: SSE-formatted response chunks. First chunk is a "start" event
+        containing the task_id for reconnection.
 
     """
-    session = await _validate_and_get_session(session_id, user_id)
+    import asyncio
+    import time
 
+    stream_start_time = time.perf_counter()
+    log_meta = {"component": "ChatStream", "session_id": session_id}
+    if user_id:
+        log_meta["user_id"] = user_id
+
+    logger.info(
+        f"[TIMING] stream_chat_post STARTED, session={session_id}, "
+        f"user={user_id}, message_len={len(request.message)}",
+        extra={"json_fields": log_meta},
+    )
+
+    _session = await _validate_and_get_session(session_id, user_id)  # noqa: F841
+    logger.info(
+        f"[TIMING] session validated in {(time.perf_counter() - stream_start_time)*1000:.1f}ms",
+        extra={
+            "json_fields": {
+                **log_meta,
+                "duration_ms": (time.perf_counter() - stream_start_time) * 1000,
+            }
+        },
+    )
+
+    # Create a task in the stream registry for reconnection support
+    task_id = str(uuid_module.uuid4())
+    operation_id = str(uuid_module.uuid4())
+    log_meta["task_id"] = task_id
+
+    task_create_start = time.perf_counter()
+    await stream_registry.create_task(
+        task_id=task_id,
+        session_id=session_id,
+        user_id=user_id,
+        tool_call_id="chat_stream",  # Not a tool call, but needed for the model
+        tool_name="chat",
+        operation_id=operation_id,
+    )
+    logger.info(
+        f"[TIMING] create_task completed in {(time.perf_counter() - task_create_start)*1000:.1f}ms",
+        extra={
+            "json_fields": {
+                **log_meta,
+                "duration_ms": (time.perf_counter() - task_create_start) * 1000,
+            }
+        },
+    )
+
+    # Enqueue the task to RabbitMQ for processing by the CoPilot executor
+    await enqueue_copilot_task(
+        task_id=task_id,
+        session_id=session_id,
+        user_id=user_id,
+        operation_id=operation_id,
+        message=request.message,
+        is_user_message=request.is_user_message,
+        context=request.context,
+    )
+
+    setup_time = (time.perf_counter() - stream_start_time) * 1000
+    logger.info(
+        f"[TIMING] Task enqueued to RabbitMQ, setup={setup_time:.1f}ms",
+        extra={"json_fields": {**log_meta, "setup_time_ms": setup_time}},
+    )
+
+    # SSE endpoint that subscribes to the task's stream
     async def event_generator() -> AsyncGenerator[str, None]:
-        chunk_count = 0
-        first_chunk_type: str | None = None
-        async for chunk in chat_service.stream_chat_completion(
-            session_id,
-            request.message,
-            is_user_message=request.is_user_message,
-            user_id=user_id,
-            session=session,  # Pass pre-fetched session to avoid double-fetch
-            context=request.context,
-        ):
-            if chunk_count < 3:
-                logger.info(
-                    "Chat stream chunk",
-                    extra={
-                        "session_id": session_id,
-                        "chunk_type": str(chunk.type),
-                    },
-                )
-            if not first_chunk_type:
-                first_chunk_type = str(chunk.type)
-            chunk_count += 1
-            yield chunk.to_sse()
+        import time as time_module
+
+        event_gen_start = time_module.perf_counter()
         logger.info(
-            "Chat stream completed",
-            extra={
-                "session_id": session_id,
-                "chunk_count": chunk_count,
-                "first_chunk_type": first_chunk_type,
-            },
+            f"[TIMING] event_generator STARTED, task={task_id}, session={session_id}, "
+            f"user={user_id}",
+            extra={"json_fields": log_meta},
         )
-        # AI SDK protocol termination
-        yield "data: [DONE]\n\n"
+        subscriber_queue = None
+        first_chunk_yielded = False
+        chunks_yielded = 0
+        try:
+            # Subscribe to the task stream (this replays existing messages + live updates)
+            subscriber_queue = await stream_registry.subscribe_to_task(
+                task_id=task_id,
+                user_id=user_id,
+                last_message_id="0-0",  # Get all messages from the beginning
+            )
+
+            if subscriber_queue is None:
+                yield StreamFinish().to_sse()
+                yield "data: [DONE]\n\n"
+                return
+
+            # Read from the subscriber queue and yield to SSE
+            logger.info(
+                "[TIMING] Starting to read from subscriber_queue",
+                extra={"json_fields": log_meta},
+            )
+            while True:
+                try:
+                    chunk = await asyncio.wait_for(subscriber_queue.get(), timeout=30.0)
+                    chunks_yielded += 1
+
+                    if not first_chunk_yielded:
+                        first_chunk_yielded = True
+                        elapsed = time_module.perf_counter() - event_gen_start
+                        logger.info(
+                            f"[TIMING] FIRST CHUNK from queue at {elapsed:.2f}s, "
+                            f"type={type(chunk).__name__}",
+                            extra={
+                                "json_fields": {
+                                    **log_meta,
+                                    "chunk_type": type(chunk).__name__,
+                                    "elapsed_ms": elapsed * 1000,
+                                }
+                            },
+                        )
+
+                    yield chunk.to_sse()
+
+                    # Check for finish signal
+                    if isinstance(chunk, StreamFinish):
+                        total_time = time_module.perf_counter() - event_gen_start
+                        logger.info(
+                            f"[TIMING] StreamFinish received in {total_time:.2f}s; "
+                            f"n_chunks={chunks_yielded}",
+                            extra={
+                                "json_fields": {
+                                    **log_meta,
+                                    "chunks_yielded": chunks_yielded,
+                                    "total_time_ms": total_time * 1000,
+                                }
+                            },
+                        )
+                        break
+                except asyncio.TimeoutError:
+                    yield StreamHeartbeat().to_sse()
+
+        except GeneratorExit:
+            logger.info(
+                f"[TIMING] GeneratorExit (client disconnected), chunks={chunks_yielded}",
+                extra={
+                    "json_fields": {
+                        **log_meta,
+                        "chunks_yielded": chunks_yielded,
+                        "reason": "client_disconnect",
+                    }
+                },
+            )
+            pass  # Client disconnected - background task continues
+        except Exception as e:
+            elapsed = (time_module.perf_counter() - event_gen_start) * 1000
+            logger.error(
+                f"[TIMING] event_generator ERROR after {elapsed:.1f}ms: {e}",
+                extra={
+                    "json_fields": {**log_meta, "elapsed_ms": elapsed, "error": str(e)}
+                },
+            )
+        finally:
+            # Unsubscribe when client disconnects or stream ends to prevent resource leak
+            if subscriber_queue is not None:
+                try:
+                    await stream_registry.unsubscribe_from_task(
+                        task_id, subscriber_queue
+                    )
+                except Exception as unsub_err:
+                    logger.error(
+                        f"Error unsubscribing from task {task_id}: {unsub_err}",
+                        exc_info=True,
+                    )
+            # AI SDK protocol termination - always yield even if unsubscribe fails
+            total_time = time_module.perf_counter() - event_gen_start
+            logger.info(
+                f"[TIMING] event_generator FINISHED in {total_time:.2f}s; "
+                f"task={task_id}, session={session_id}, n_chunks={chunks_yielded}",
+                extra={
+                    "json_fields": {
+                        **log_meta,
+                        "total_time_ms": total_time * 1000,
+                        "chunks_yielded": chunks_yielded,
+                    }
+                },
+            )
+            yield "data: [DONE]\n\n"
 
     return StreamingResponse(
         event_generator(),
@@ -270,63 +494,90 @@ async def stream_chat_post(
 @router.get(
     "/sessions/{session_id}/stream",
 )
-async def stream_chat_get(
+async def resume_session_stream(
     session_id: str,
-    message: Annotated[str, Query(min_length=1, max_length=10000)],
     user_id: str | None = Depends(auth.get_user_id),
-    is_user_message: bool = Query(default=True),
 ):
     """
-    Stream chat responses for a session (GET - legacy endpoint).
+    Resume an active stream for a session.
 
-    Streams the AI/completion responses in real time over Server-Sent Events (SSE), including:
-      - Text fragments as they are generated
-      - Tool call UI elements (if invoked)
-      - Tool execution results
+    Called by the AI SDK's ``useChat(resume: true)`` on page load.
+    Checks for an active (in-progress) task on the session and either replays
+    the full SSE stream or returns 204 No Content if nothing is running.
 
     Args:
-        session_id: The chat session identifier to associate with the streamed messages.
-        message: The user's new message to process.
+        session_id: The chat session identifier.
         user_id: Optional authenticated user ID.
-        is_user_message: Whether the message is a user message.
-    Returns:
-        StreamingResponse: SSE-formatted response chunks.
 
+    Returns:
+        StreamingResponse (SSE) when an active stream exists,
+        or 204 No Content when there is nothing to resume.
     """
-    session = await _validate_and_get_session(session_id, user_id)
+    import asyncio
+
+    active_task, _last_id = await stream_registry.get_active_task_for_session(
+        session_id, user_id
+    )
+
+    if not active_task:
+        return Response(status_code=204)
+
+    subscriber_queue = await stream_registry.subscribe_to_task(
+        task_id=active_task.task_id,
+        user_id=user_id,
+        last_message_id="0-0",  # Full replay so useChat rebuilds the message
+    )
+
+    if subscriber_queue is None:
+        return Response(status_code=204)
 
     async def event_generator() -> AsyncGenerator[str, None]:
         chunk_count = 0
         first_chunk_type: str | None = None
-        async for chunk in chat_service.stream_chat_completion(
-            session_id,
-            message,
-            is_user_message=is_user_message,
-            user_id=user_id,
-            session=session,  # Pass pre-fetched session to avoid double-fetch
-        ):
-            if chunk_count < 3:
-                logger.info(
-                    "Chat stream chunk",
-                    extra={
-                        "session_id": session_id,
-                        "chunk_type": str(chunk.type),
-                    },
+        try:
+            while True:
+                try:
+                    chunk = await asyncio.wait_for(subscriber_queue.get(), timeout=30.0)
+                    if chunk_count < 3:
+                        logger.info(
+                            "Resume stream chunk",
+                            extra={
+                                "session_id": session_id,
+                                "chunk_type": str(chunk.type),
+                            },
+                        )
+                    if not first_chunk_type:
+                        first_chunk_type = str(chunk.type)
+                    chunk_count += 1
+                    yield chunk.to_sse()
+
+                    if isinstance(chunk, StreamFinish):
+                        break
+                except asyncio.TimeoutError:
+                    yield StreamHeartbeat().to_sse()
+        except GeneratorExit:
+            pass
+        except Exception as e:
+            logger.error(f"Error in resume stream for session {session_id}: {e}")
+        finally:
+            try:
+                await stream_registry.unsubscribe_from_task(
+                    active_task.task_id, subscriber_queue
                 )
-            if not first_chunk_type:
-                first_chunk_type = str(chunk.type)
-            chunk_count += 1
-            yield chunk.to_sse()
-        logger.info(
-            "Chat stream completed",
-            extra={
-                "session_id": session_id,
-                "chunk_count": chunk_count,
-                "first_chunk_type": first_chunk_type,
-            },
-        )
-        # AI SDK protocol termination
-        yield "data: [DONE]\n\n"
+            except Exception as unsub_err:
+                logger.error(
+                    f"Error unsubscribing from task {active_task.task_id}: {unsub_err}",
+                    exc_info=True,
+                )
+            logger.info(
+                "Resume stream completed",
+                extra={
+                    "session_id": session_id,
+                    "n_chunks": chunk_count,
+                    "first_chunk_type": first_chunk_type,
+                },
+            )
+            yield "data: [DONE]\n\n"
 
     return StreamingResponse(
         event_generator(),
@@ -334,8 +585,8 @@ async def stream_chat_get(
         headers={
             "Cache-Control": "no-cache",
             "Connection": "keep-alive",
-            "X-Accel-Buffering": "no",  # Disable nginx buffering
-            "x-vercel-ai-ui-message-stream": "v1",  # AI SDK protocol header
+            "X-Accel-Buffering": "no",
+            "x-vercel-ai-ui-message-stream": "v1",
         },
     )
 
@@ -366,6 +617,251 @@ async def session_assign_user(
     return {"status": "ok"}
 
 
+# ========== Task Streaming (SSE Reconnection) ==========
+
+
+@router.get(
+    "/tasks/{task_id}/stream",
+)
+async def stream_task(
+    task_id: str,
+    user_id: str | None = Depends(auth.get_user_id),
+    last_message_id: str = Query(
+        default="0-0",
+        description="Last Redis Stream message ID received (e.g., '1706540123456-0'). Use '0-0' for full replay.",
+    ),
+):
+    """
+    Reconnect to a long-running task's SSE stream.
+
+    When a long-running operation (like agent generation) starts, the client
+    receives a task_id. If the connection drops, the client can reconnect
+    using this endpoint to resume receiving updates.
+
+    Args:
+        task_id: The task ID from the operation_started response.
+        user_id: Authenticated user ID for ownership validation.
+        last_message_id: Last Redis Stream message ID received ("0-0" for full replay).
+
+    Returns:
+        StreamingResponse: SSE-formatted response chunks starting after last_message_id.
+
+    Raises:
+        HTTPException: 404 if task not found, 410 if task expired, 403 if access denied.
+    """
+    # Check task existence and expiry before subscribing
+    task, error_code = await stream_registry.get_task_with_expiry_info(task_id)
+
+    if error_code == "TASK_EXPIRED":
+        raise HTTPException(
+            status_code=410,
+            detail={
+                "code": "TASK_EXPIRED",
+                "message": "This operation has expired. Please try again.",
+            },
+        )
+
+    if error_code == "TASK_NOT_FOUND":
+        raise HTTPException(
+            status_code=404,
+            detail={
+                "code": "TASK_NOT_FOUND",
+                "message": f"Task {task_id} not found.",
+            },
+        )
+
+    # Validate ownership if task has an owner
+    if task and task.user_id and user_id != task.user_id:
+        raise HTTPException(
+            status_code=403,
+            detail={
+                "code": "ACCESS_DENIED",
+                "message": "You do not have access to this task.",
+            },
+        )
+
+    # Get subscriber queue from stream registry
+    subscriber_queue = await stream_registry.subscribe_to_task(
+        task_id=task_id,
+        user_id=user_id,
+        last_message_id=last_message_id,
+    )
+
+    if subscriber_queue is None:
+        raise HTTPException(
+            status_code=404,
+            detail={
+                "code": "TASK_NOT_FOUND",
+                "message": f"Task {task_id} not found or access denied.",
+            },
+        )
+
+    async def event_generator() -> AsyncGenerator[str, None]:
+        import asyncio
+
+        heartbeat_interval = 15.0  # Send heartbeat every 15 seconds
+        try:
+            while True:
+                try:
+                    # Wait for next chunk with timeout for heartbeats
+                    chunk = await asyncio.wait_for(
+                        subscriber_queue.get(), timeout=heartbeat_interval
+                    )
+                    yield chunk.to_sse()
+
+                    # Check for finish signal
+                    if isinstance(chunk, StreamFinish):
+                        break
+                except asyncio.TimeoutError:
+                    # Send heartbeat to keep connection alive
+                    yield StreamHeartbeat().to_sse()
+        except Exception as e:
+            logger.error(f"Error in task stream {task_id}: {e}", exc_info=True)
+        finally:
+            # Unsubscribe when client disconnects or stream ends
+            try:
+                await stream_registry.unsubscribe_from_task(task_id, subscriber_queue)
+            except Exception as unsub_err:
+                logger.error(
+                    f"Error unsubscribing from task {task_id}: {unsub_err}",
+                    exc_info=True,
+                )
+            # AI SDK protocol termination - always yield even if unsubscribe fails
+            yield "data: [DONE]\n\n"
+
+    return StreamingResponse(
+        event_generator(),
+        media_type="text/event-stream",
+        headers={
+            "Cache-Control": "no-cache",
+            "Connection": "keep-alive",
+            "X-Accel-Buffering": "no",
+            "x-vercel-ai-ui-message-stream": "v1",
+        },
+    )
+
+
+@router.get(
+    "/tasks/{task_id}",
+)
+async def get_task_status(
+    task_id: str,
+    user_id: str | None = Depends(auth.get_user_id),
+) -> dict:
+    """
+    Get the status of a long-running task.
+
+    Args:
+        task_id: The task ID to check.
+        user_id: Authenticated user ID for ownership validation.
+
+    Returns:
+        dict: Task status including task_id, status, tool_name, and operation_id.
+
+    Raises:
+        NotFoundError: If task_id is not found or user doesn't have access.
+    """
+    task = await stream_registry.get_task(task_id)
+
+    if task is None:
+        raise NotFoundError(f"Task {task_id} not found.")
+
+    # Validate ownership - if task has an owner, requester must match
+    if task.user_id and user_id != task.user_id:
+        raise NotFoundError(f"Task {task_id} not found.")
+
+    return {
+        "task_id": task.task_id,
+        "session_id": task.session_id,
+        "status": task.status,
+        "tool_name": task.tool_name,
+        "operation_id": task.operation_id,
+        "created_at": task.created_at.isoformat(),
+    }
+
+
+# ========== External Completion Webhook ==========
+
+
+@router.post(
+    "/operations/{operation_id}/complete",
+    status_code=200,
+)
+async def complete_operation(
+    operation_id: str,
+    request: OperationCompleteRequest,
+    x_api_key: str | None = Header(default=None),
+) -> dict:
+    """
+    External completion webhook for long-running operations.
+
+    Called by Agent Generator (or other services) when an operation completes.
+    This triggers the stream registry to publish completion and continue LLM generation.
+
+    Args:
+        operation_id: The operation ID to complete.
+        request: Completion payload with success status and result/error.
+        x_api_key: Internal API key for authentication.
+
+    Returns:
+        dict: Status of the completion.
+
+    Raises:
+        HTTPException: If API key is invalid or operation not found.
+    """
+    # Validate internal API key - reject if not configured or invalid
+    if not config.internal_api_key:
+        logger.error(
+            "Operation complete webhook rejected: CHAT_INTERNAL_API_KEY not configured"
+        )
+        raise HTTPException(
+            status_code=503,
+            detail="Webhook not available: internal API key not configured",
+        )
+    if x_api_key != config.internal_api_key:
+        raise HTTPException(status_code=401, detail="Invalid API key")
+
+    # Find task by operation_id
+    task = await stream_registry.find_task_by_operation_id(operation_id)
+    if task is None:
+        raise HTTPException(
+            status_code=404,
+            detail=f"Operation {operation_id} not found",
+        )
+
+    logger.info(
+        f"Received completion webhook for operation {operation_id} "
+        f"(task_id={task.task_id}, success={request.success})"
+    )
+
+    if request.success:
+        await process_operation_success(task, request.result)
+    else:
+        await process_operation_failure(task, request.error)
+
+    return {"status": "ok", "task_id": task.task_id}
+
+
+# ========== Configuration ==========
+
+
+@router.get("/config/ttl", status_code=200)
+async def get_ttl_config() -> dict:
+    """
+    Get the stream TTL configuration.
+
+    Returns the Time-To-Live settings for chat streams, which determines
+    how long clients can reconnect to an active stream.
+
+    Returns:
+        dict: TTL configuration with seconds and milliseconds values.
+    """
+    return {
+        "stream_ttl_seconds": config.stream_ttl,
+        "stream_ttl_ms": config.stream_ttl * 1000,
+    }
+
+
 # ========== Health Check ==========
 
 
@@ -402,3 +898,42 @@ async def health_check() -> dict:
         "service": "chat",
         "version": "0.1.0",
     }
+
+
+# ========== Schema Export (for OpenAPI / Orval codegen) ==========
+
+ToolResponseUnion = (
+    AgentsFoundResponse
+    | NoResultsResponse
+    | AgentDetailsResponse
+    | SetupRequirementsResponse
+    | ExecutionStartedResponse
+    | NeedLoginResponse
+    | ErrorResponse
+    | InputValidationErrorResponse
+    | AgentOutputResponse
+    | UnderstandingUpdatedResponse
+    | AgentPreviewResponse
+    | AgentSavedResponse
+    | ClarificationNeededResponse
+    | BlockListResponse
+    | BlockOutputResponse
+    | DocSearchResultsResponse
+    | DocPageResponse
+    | OperationStartedResponse
+    | OperationPendingResponse
+    | OperationInProgressResponse
+)
+
+
+@router.get(
+    "/schema/tool-responses",
+    response_model=ToolResponseUnion,
+    include_in_schema=True,
+    summary="[Dummy] Tool response type export for codegen",
+    description="This endpoint is not meant to be called. It exists solely to "
+    "expose tool response models in the OpenAPI schema for frontend codegen.",
+)
+async def _tool_response_schema() -> ToolResponseUnion:  # type: ignore[return]
+    """Never called at runtime. Exists only so Orval generates TS types."""
+    raise HTTPException(status_code=501, detail="Schema-only endpoint")

autogpt_platform/backend/backend/api/features/chat/service.py

@@ -1,910 +0,0 @@
-import asyncio
-import logging
-import time
-from asyncio import CancelledError
-from collections.abc import AsyncGenerator
-from typing import Any
-
-import orjson
-from langfuse import get_client, propagate_attributes
-from langfuse.openai import openai  # type: ignore
-from openai import (
-    APIConnectionError,
-    APIError,
-    APIStatusError,
-    PermissionDeniedError,
-    RateLimitError,
-)
-from openai.types.chat import ChatCompletionChunk, ChatCompletionToolParam
-
-from backend.data.understanding import (
-    format_understanding_for_prompt,
-    get_business_understanding,
-)
-from backend.util.exceptions import NotFoundError
-from backend.util.settings import Settings
-
-from .config import ChatConfig
-from .model import (
-    ChatMessage,
-    ChatSession,
-    Usage,
-    cache_chat_session,
-    get_chat_session,
-    update_session_title,
-    upsert_chat_session,
-)
-from .response_model import (
-    StreamBaseResponse,
-    StreamError,
-    StreamFinish,
-    StreamStart,
-    StreamTextDelta,
-    StreamTextEnd,
-    StreamTextStart,
-    StreamToolInputAvailable,
-    StreamToolInputStart,
-    StreamToolOutputAvailable,
-    StreamUsage,
-)
-from .tools import execute_tool, tools
-
-logger = logging.getLogger(__name__)
-
-config = ChatConfig()
-settings = Settings()
-client = openai.AsyncOpenAI(api_key=config.api_key, base_url=config.base_url)
-
-
-langfuse = get_client()
-
-
-class LangfuseNotConfiguredError(Exception):
-    """Raised when Langfuse is required but not configured."""
-
-    pass
-
-
-def _is_langfuse_configured() -> bool:
-    """Check if Langfuse credentials are configured."""
-    return bool(
-        settings.secrets.langfuse_public_key and settings.secrets.langfuse_secret_key
-    )
-
-
-async def _build_system_prompt(user_id: str | None) -> tuple[str, Any]:
-    """Build the full system prompt including business understanding if available.
-
-    Args:
-        user_id: The user ID for fetching business understanding
-                     If "default" and this is the user's first session, will use "onboarding" instead.
-
-    Returns:
-        Tuple of (compiled prompt string, Langfuse prompt object for tracing)
-    """
-
-    # cache_ttl_seconds=0 disables SDK caching to always get the latest prompt
-    prompt = langfuse.get_prompt(config.langfuse_prompt_name, cache_ttl_seconds=0)
-
-    # If user is authenticated, try to fetch their business understanding
-    understanding = None
-    if user_id:
-        try:
-            understanding = await get_business_understanding(user_id)
-        except Exception as e:
-            logger.warning(f"Failed to fetch business understanding: {e}")
-            understanding = None
-    if understanding:
-        context = format_understanding_for_prompt(understanding)
-    else:
-        context = "This is the first time you are meeting the user. Greet them and introduce them to the platform"
-
-    compiled = prompt.compile(users_information=context)
-    return compiled, understanding
-
-
-async def _generate_session_title(message: str) -> str | None:
-    """Generate a concise title for a chat session based on the first message.
-
-    Args:
-        message: The first user message in the session
-
-    Returns:
-        A short title (3-6 words) or None if generation fails
-    """
-    try:
-        response = await client.chat.completions.create(
-            model=config.title_model,
-            messages=[
-                {
-                    "role": "system",
-                    "content": (
-                        "Generate a very short title (3-6 words) for a chat conversation "
-                        "based on the user's first message. The title should capture the "
-                        "main topic or intent. Return ONLY the title, no quotes or punctuation."
-                    ),
-                },
-                {"role": "user", "content": message[:500]},  # Limit input length
-            ],
-            max_tokens=20,
-        )
-        title = response.choices[0].message.content
-        if title:
-            # Clean up the title
-            title = title.strip().strip("\"'")
-            # Limit length
-            if len(title) > 50:
-                title = title[:47] + "..."
-            return title
-        return None
-    except Exception as e:
-        logger.warning(f"Failed to generate session title: {e}")
-        return None
-
-
-async def assign_user_to_session(
-    session_id: str,
-    user_id: str,
-) -> ChatSession:
-    """
-    Assign a user to a chat session.
-    """
-    session = await get_chat_session(session_id, None)
-    if not session:
-        raise NotFoundError(f"Session {session_id} not found")
-    session.user_id = user_id
-    return await upsert_chat_session(session)
-
-
-async def stream_chat_completion(
-    session_id: str,
-    message: str | None = None,
-    tool_call_response: str | None = None,
-    is_user_message: bool = True,
-    user_id: str | None = None,
-    retry_count: int = 0,
-    session: ChatSession | None = None,
-    context: dict[str, str] | None = None,  # {url: str, content: str}
-) -> AsyncGenerator[StreamBaseResponse, None]:
-    """Main entry point for streaming chat completions with database handling.
-
-    This function handles all database operations and delegates streaming
-    to the internal _stream_chat_chunks function.
-
-    Args:
-        session_id: Chat session ID
-        user_message: User's input message
-        user_id: User ID for authentication (None for anonymous)
-        session: Optional pre-loaded session object (for recursive calls to avoid Redis refetch)
-
-    Yields:
-        StreamBaseResponse objects formatted as SSE
-
-    Raises:
-        NotFoundError: If session_id is invalid
-        ValueError: If max_context_messages is exceeded
-
-    """
-    logger.info(
-        f"Streaming chat completion for session {session_id} for message {message} and user id {user_id}. Message is user message: {is_user_message}"
-    )
-
-    # Check if Langfuse is configured - required for chat functionality
-    if not _is_langfuse_configured():
-        logger.error("Chat request failed: Langfuse is not configured")
-        yield StreamError(
-            errorText="Chat service is not available. Langfuse must be configured "
-            "with LANGFUSE_PUBLIC_KEY and LANGFUSE_SECRET_KEY environment variables."
-        )
-        yield StreamFinish()
-        return
-
-    # Only fetch from Redis if session not provided (initial call)
-    if session is None:
-        session = await get_chat_session(session_id, user_id)
-        logger.info(
-            f"Fetched session from Redis: {session.session_id if session else 'None'}, "
-            f"message_count={len(session.messages) if session else 0}"
-        )
-    else:
-        logger.info(
-            f"Using provided session object: {session.session_id}, "
-            f"message_count={len(session.messages)}"
-        )
-
-    if not session:
-        raise NotFoundError(
-            f"Session {session_id} not found. Please create a new session first."
-        )
-
-    if message:
-        # Build message content with context if provided
-        message_content = message
-        if context and context.get("url") and context.get("content"):
-            context_text = f"Page URL: {context['url']}\n\nPage Content:\n{context['content']}\n\n---\n\nUser Message: {message}"
-            message_content = context_text
-            logger.info(
-                f"Including page context: URL={context['url']}, content_length={len(context['content'])}"
-            )
-
-        session.messages.append(
-            ChatMessage(
-                role="user" if is_user_message else "assistant", content=message_content
-            )
-        )
-        logger.info(
-            f"Appended message (role={'user' if is_user_message else 'assistant'}), "
-            f"new message_count={len(session.messages)}"
-        )
-
-    logger.info(
-        f"Upserting session: {session.session_id} with user id {session.user_id}, "
-        f"message_count={len(session.messages)}"
-    )
-    session = await upsert_chat_session(session)
-    assert session, "Session not found"
-
-    # Generate title for new sessions on first user message (non-blocking)
-    # Check: is_user_message, no title yet, and this is the first user message
-    if is_user_message and message and not session.title:
-        user_messages = [m for m in session.messages if m.role == "user"]
-        if len(user_messages) == 1:
-            # First user message - generate title in background
-            import asyncio
-
-            # Capture only the values we need (not the session object) to avoid
-            # stale data issues when the main flow modifies the session
-            captured_session_id = session_id
-            captured_message = message
-
-            async def _update_title():
-                try:
-                    title = await _generate_session_title(captured_message)
-                    if title:
-                        # Use dedicated title update function that doesn't
-                        # touch messages, avoiding race conditions
-                        await update_session_title(captured_session_id, title)
-                        logger.info(
-                            f"Generated title for session {captured_session_id}: {title}"
-                        )
-                except Exception as e:
-                    logger.warning(f"Failed to update session title: {e}")
-
-            # Fire and forget - don't block the chat response
-            asyncio.create_task(_update_title())
-
-    # Build system prompt with business understanding
-    system_prompt, understanding = await _build_system_prompt(user_id)
-
-    # Create Langfuse trace for this LLM call (each call gets its own trace, grouped by session_id)
-    # Using v3 SDK: start_observation creates a root span, update_trace sets trace-level attributes
-    input = message
-    if not message and tool_call_response:
-        input = tool_call_response
-
-    langfuse = get_client()
-    with langfuse.start_as_current_observation(
-        as_type="span",
-        name="user-copilot-request",
-        input=input,
-    ) as span:
-        with propagate_attributes(
-            session_id=session_id,
-            user_id=user_id,
-            tags=["copilot"],
-            metadata={
-                "users_information": format_understanding_for_prompt(understanding)[
-                    :200
-                ]  # langfuse only accepts upto to 200 chars
-            },
-        ):
-
-            # Initialize variables that will be used in finally block (must be defined before try)
-            assistant_response = ChatMessage(
-                role="assistant",
-                content="",
-            )
-            accumulated_tool_calls: list[dict[str, Any]] = []
-            has_saved_assistant_message = False
-            has_appended_streaming_message = False
-            last_cache_time = 0.0
-            last_cache_content_len = 0
-
-            # Wrap main logic in try/finally to ensure Langfuse observations are always ended
-            has_yielded_end = False
-            has_yielded_error = False
-            has_done_tool_call = False
-            has_received_text = False
-            text_streaming_ended = False
-            tool_response_messages: list[ChatMessage] = []
-            should_retry = False
-
-            # Generate unique IDs for AI SDK protocol
-            import uuid as uuid_module
-
-            message_id = str(uuid_module.uuid4())
-            text_block_id = str(uuid_module.uuid4())
-
-            # Yield message start
-            yield StreamStart(messageId=message_id)
-
-            try:
-                async for chunk in _stream_chat_chunks(
-                    session=session,
-                    tools=tools,
-                    system_prompt=system_prompt,
-                    text_block_id=text_block_id,
-                ):
-
-                    if isinstance(chunk, StreamTextStart):
-                        # Emit text-start before first text delta
-                        if not has_received_text:
-                            yield chunk
-                    elif isinstance(chunk, StreamTextDelta):
-                        delta = chunk.delta or ""
-                        assert assistant_response.content is not None
-                        assistant_response.content += delta
-                        has_received_text = True
-                        if not has_appended_streaming_message:
-                            session.messages.append(assistant_response)
-                            has_appended_streaming_message = True
-                        current_time = time.monotonic()
-                        content_len = len(assistant_response.content)
-                        if (
-                            current_time - last_cache_time >= 1.0
-                            and content_len > last_cache_content_len
-                        ):
-                            try:
-                                await cache_chat_session(session)
-                            except Exception as e:
-                                logger.warning(
-                                    f"Failed to cache partial session {session.session_id}: {e}"
-                                )
-                            last_cache_time = current_time
-                            last_cache_content_len = content_len
-                        yield chunk
-                    elif isinstance(chunk, StreamTextEnd):
-                        # Emit text-end after text completes
-                        if has_received_text and not text_streaming_ended:
-                            text_streaming_ended = True
-                            if assistant_response.content:
-                                logger.warn(
-                                    f"StreamTextEnd: Attempting to set output {assistant_response.content}"
-                                )
-                                span.update_trace(output=assistant_response.content)
-                                span.update(output=assistant_response.content)
-                            yield chunk
-                    elif isinstance(chunk, StreamToolInputStart):
-                        # Emit text-end before first tool call, but only if we've received text
-                        if has_received_text and not text_streaming_ended:
-                            yield StreamTextEnd(id=text_block_id)
-                            text_streaming_ended = True
-                        yield chunk
-                    elif isinstance(chunk, StreamToolInputAvailable):
-                        # Accumulate tool calls in OpenAI format
-                        accumulated_tool_calls.append(
-                            {
-                                "id": chunk.toolCallId,
-                                "type": "function",
-                                "function": {
-                                    "name": chunk.toolName,
-                                    "arguments": orjson.dumps(chunk.input).decode(
-                                        "utf-8"
-                                    ),
-                                },
-                            }
-                        )
-                    elif isinstance(chunk, StreamToolOutputAvailable):
-                        result_content = (
-                            chunk.output
-                            if isinstance(chunk.output, str)
-                            else orjson.dumps(chunk.output).decode("utf-8")
-                        )
-                        tool_response_messages.append(
-                            ChatMessage(
-                                role="tool",
-                                content=result_content,
-                                tool_call_id=chunk.toolCallId,
-                            )
-                        )
-                        has_done_tool_call = True
-                        # Track if any tool execution failed
-                        if not chunk.success:
-                            logger.warning(
-                                f"Tool {chunk.toolName} (ID: {chunk.toolCallId}) execution failed"
-                            )
-                        yield chunk
-                    elif isinstance(chunk, StreamFinish):
-                        if not has_done_tool_call:
-                            # Emit text-end before finish if we received text but haven't closed it
-                            if has_received_text and not text_streaming_ended:
-                                yield StreamTextEnd(id=text_block_id)
-                                text_streaming_ended = True
-
-                            # Save assistant message before yielding finish to ensure it's persisted
-                            # even if client disconnects immediately after receiving StreamFinish
-                            if not has_saved_assistant_message:
-                                messages_to_save_early: list[ChatMessage] = []
-                                if accumulated_tool_calls:
-                                    assistant_response.tool_calls = (
-                                        accumulated_tool_calls
-                                    )
-                                if not has_appended_streaming_message and (
-                                    assistant_response.content
-                                    or assistant_response.tool_calls
-                                ):
-                                    messages_to_save_early.append(assistant_response)
-                                messages_to_save_early.extend(tool_response_messages)
-
-                                if messages_to_save_early:
-                                    session.messages.extend(messages_to_save_early)
-                                    logger.info(
-                                        f"Saving assistant message before StreamFinish: "
-                                        f"content_len={len(assistant_response.content or '')}, "
-                                        f"tool_calls={len(assistant_response.tool_calls or [])}, "
-                                        f"tool_responses={len(tool_response_messages)}"
-                                    )
-                                if (
-                                    messages_to_save_early
-                                    or has_appended_streaming_message
-                                ):
-                                    await upsert_chat_session(session)
-                                    has_saved_assistant_message = True
-
-                            has_yielded_end = True
-                            yield chunk
-                    elif isinstance(chunk, StreamError):
-                        has_yielded_error = True
-                        yield chunk
-                    elif isinstance(chunk, StreamUsage):
-                        session.usage.append(
-                            Usage(
-                                prompt_tokens=chunk.promptTokens,
-                                completion_tokens=chunk.completionTokens,
-                                total_tokens=chunk.totalTokens,
-                            )
-                        )
-                    else:
-                        logger.error(
-                            f"Unknown chunk type: {type(chunk)}", exc_info=True
-                        )
-                if assistant_response.content:
-                    langfuse.update_current_trace(output=assistant_response.content)
-                    langfuse.update_current_span(output=assistant_response.content)
-                elif tool_response_messages:
-                    langfuse.update_current_trace(output=str(tool_response_messages))
-                    langfuse.update_current_span(output=str(tool_response_messages))
-
-            except CancelledError:
-                if not has_saved_assistant_message:
-                    if accumulated_tool_calls:
-                        assistant_response.tool_calls = accumulated_tool_calls
-                    if assistant_response.content:
-                        assistant_response.content = (
-                            f"{assistant_response.content}\n\n[interrupted]"
-                        )
-                    else:
-                        assistant_response.content = "[interrupted]"
-                    if not has_appended_streaming_message:
-                        session.messages.append(assistant_response)
-                    if tool_response_messages:
-                        session.messages.extend(tool_response_messages)
-                    try:
-                        await upsert_chat_session(session)
-                    except Exception as e:
-                        logger.warning(
-                            f"Failed to save interrupted session {session.session_id}: {e}"
-                        )
-                raise
-            except Exception as e:
-                logger.error(f"Error during stream: {e!s}", exc_info=True)
-
-                # Check if this is a retryable error (JSON parsing, incomplete tool calls, etc.)
-                is_retryable = isinstance(
-                    e, (orjson.JSONDecodeError, KeyError, TypeError)
-                )
-
-                if is_retryable and retry_count < config.max_retries:
-                    logger.info(
-                        f"Retryable error encountered. Attempt {retry_count + 1}/{config.max_retries}"
-                    )
-                    should_retry = True
-                else:
-                    # Non-retryable error or max retries exceeded
-                    # Save any partial progress before reporting error
-                    messages_to_save: list[ChatMessage] = []
-
-                    # Add assistant message if it has content or tool calls
-                    if accumulated_tool_calls:
-                        assistant_response.tool_calls = accumulated_tool_calls
-                    if not has_appended_streaming_message and (
-                        assistant_response.content or assistant_response.tool_calls
-                    ):
-                        messages_to_save.append(assistant_response)
-
-                    # Add tool response messages after assistant message
-                    messages_to_save.extend(tool_response_messages)
-
-                    if not has_saved_assistant_message:
-                        if messages_to_save:
-                            session.messages.extend(messages_to_save)
-                        if messages_to_save or has_appended_streaming_message:
-                            await upsert_chat_session(session)
-
-                    if not has_yielded_error:
-                        error_message = str(e)
-                        if not is_retryable:
-                            error_message = f"Non-retryable error: {error_message}"
-                        elif retry_count >= config.max_retries:
-                            error_message = f"Max retries ({config.max_retries}) exceeded: {error_message}"
-
-                        error_response = StreamError(errorText=error_message)
-                        yield error_response
-                    if not has_yielded_end:
-                        yield StreamFinish()
-                    return
-
-            # Handle retry outside of exception handler to avoid nesting
-            if should_retry and retry_count < config.max_retries:
-                logger.info(
-                    f"Retrying stream_chat_completion for session {session_id}, attempt {retry_count + 1}"
-                )
-                async for chunk in stream_chat_completion(
-                    session_id=session.session_id,
-                    user_id=user_id,
-                    retry_count=retry_count + 1,
-                    session=session,
-                    context=context,
-                ):
-                    yield chunk
-                return  # Exit after retry to avoid double-saving in finally block
-
-            # Normal completion path - save session and handle tool call continuation
-            # Only save if we haven't already saved when StreamFinish was received
-            if not has_saved_assistant_message:
-                logger.info(
-                    f"Normal completion path: session={session.session_id}, "
-                    f"current message_count={len(session.messages)}"
-                )
-
-                # Build the messages list in the correct order
-                messages_to_save: list[ChatMessage] = []
-
-                # Add assistant message with tool_calls if any
-                if accumulated_tool_calls:
-                    assistant_response.tool_calls = accumulated_tool_calls
-                    logger.info(
-                        f"Added {len(accumulated_tool_calls)} tool calls to assistant message"
-                    )
-                if not has_appended_streaming_message and (
-                    assistant_response.content or assistant_response.tool_calls
-                ):
-                    messages_to_save.append(assistant_response)
-                    logger.info(
-                        f"Saving assistant message with content_len={len(assistant_response.content or '')}, tool_calls={len(assistant_response.tool_calls or [])}"
-                    )
-
-                # Add tool response messages after assistant message
-                messages_to_save.extend(tool_response_messages)
-                logger.info(
-                    f"Saving {len(tool_response_messages)} tool response messages, "
-                    f"total_to_save={len(messages_to_save)}"
-                )
-
-                if messages_to_save:
-                    session.messages.extend(messages_to_save)
-                    logger.info(
-                        f"Extended session messages, new message_count={len(session.messages)}"
-                    )
-                if messages_to_save or has_appended_streaming_message:
-                    await upsert_chat_session(session)
-            else:
-                logger.info(
-                    "Assistant message already saved when StreamFinish was received, "
-                    "skipping duplicate save"
-                )
-
-            # If we did a tool call, stream the chat completion again to get the next response
-            if has_done_tool_call:
-                logger.info(
-                    "Tool call executed, streaming chat completion again to get assistant response"
-                )
-                async for chunk in stream_chat_completion(
-                    session_id=session.session_id,
-                    user_id=user_id,
-                    session=session,  # Pass session object to avoid Redis refetch
-                    context=context,
-                    tool_call_response=str(tool_response_messages),
-                ):
-                    yield chunk
-
-
-# Retry configuration for OpenAI API calls
-MAX_RETRIES = 3
-BASE_DELAY_SECONDS = 1.0
-MAX_DELAY_SECONDS = 30.0
-
-
-def _is_retryable_error(error: Exception) -> bool:
-    """Determine if an error is retryable."""
-    if isinstance(error, RateLimitError):
-        return True
-    if isinstance(error, APIConnectionError):
-        return True
-    if isinstance(error, APIStatusError):
-        # APIStatusError has a response with status_code
-        # Retry on 5xx status codes (server errors)
-        if error.response.status_code >= 500:
-            return True
-    if isinstance(error, APIError):
-        # Retry on overloaded errors or 500 errors (may not have status code)
-        error_message = str(error).lower()
-        if "overloaded" in error_message or "internal server error" in error_message:
-            return True
-    return False
-
-
-def _is_region_blocked_error(error: Exception) -> bool:
-    if isinstance(error, PermissionDeniedError):
-        return "not available in your region" in str(error).lower()
-    return "not available in your region" in str(error).lower()
-
-
-async def _stream_chat_chunks(
-    session: ChatSession,
-    tools: list[ChatCompletionToolParam],
-    system_prompt: str | None = None,
-    text_block_id: str | None = None,
-) -> AsyncGenerator[StreamBaseResponse, None]:
-    """
-    Pure streaming function for OpenAI chat completions with tool calling.
-
-    This function is database-agnostic and focuses only on streaming logic.
-    Implements exponential backoff retry for transient API errors.
-
-    Args:
-        session: Chat session with conversation history
-        tools: Available tools for the model
-        system_prompt: System prompt to prepend to messages
-
-    Yields:
-        SSE formatted JSON response objects
-
-    """
-    model = config.model
-
-    logger.info("Starting pure chat stream")
-
-    # Build messages with system prompt prepended
-    messages = session.to_openai_messages()
-    if system_prompt:
-        from openai.types.chat import ChatCompletionSystemMessageParam
-
-        system_message = ChatCompletionSystemMessageParam(
-            role="system",
-            content=system_prompt,
-        )
-        messages = [system_message] + messages
-
-    # Loop to handle tool calls and continue conversation
-    while True:
-        retry_count = 0
-        last_error: Exception | None = None
-
-        while retry_count <= MAX_RETRIES:
-            try:
-                logger.info(
-                    f"Creating OpenAI chat completion stream..."
-                    f"{f' (retry {retry_count}/{MAX_RETRIES})' if retry_count > 0 else ''}"
-                )
-
-                # Create the stream with proper types
-                stream = await client.chat.completions.create(
-                    model=model,
-                    messages=messages,
-                    tools=tools,
-                    tool_choice="auto",
-                    stream=True,
-                    stream_options={"include_usage": True},
-                )
-
-                # Variables to accumulate tool calls
-                tool_calls: list[dict[str, Any]] = []
-                active_tool_call_idx: int | None = None
-                finish_reason: str | None = None
-                # Track which tool call indices have had their start event emitted
-                emitted_start_for_idx: set[int] = set()
-
-                # Track if we've started the text block
-                text_started = False
-
-                # Process the stream
-                chunk: ChatCompletionChunk
-                async for chunk in stream:
-                    if chunk.usage:
-                        yield StreamUsage(
-                            promptTokens=chunk.usage.prompt_tokens,
-                            completionTokens=chunk.usage.completion_tokens,
-                            totalTokens=chunk.usage.total_tokens,
-                        )
-
-                    if chunk.choices:
-                        choice = chunk.choices[0]
-                        delta = choice.delta
-
-                        # Capture finish reason
-                        if choice.finish_reason:
-                            finish_reason = choice.finish_reason
-                            logger.info(f"Finish reason: {finish_reason}")
-
-                        # Handle content streaming
-                        if delta.content:
-                            # Emit text-start on first text content
-                            if not text_started and text_block_id:
-                                yield StreamTextStart(id=text_block_id)
-                                text_started = True
-                            # Stream the text delta
-                            text_response = StreamTextDelta(
-                                id=text_block_id or "",
-                                delta=delta.content,
-                            )
-                            yield text_response
-
-                        # Handle tool calls
-                        if delta.tool_calls:
-                            for tc_chunk in delta.tool_calls:
-                                idx = tc_chunk.index
-
-                                # Update active tool call index if needed
-                                if (
-                                    active_tool_call_idx is None
-                                    or active_tool_call_idx != idx
-                                ):
-                                    active_tool_call_idx = idx
-
-                                # Ensure we have a tool call object at this index
-                                while len(tool_calls) <= idx:
-                                    tool_calls.append(
-                                        {
-                                            "id": "",
-                                            "type": "function",
-                                            "function": {
-                                                "name": "",
-                                                "arguments": "",
-                                            },
-                                        },
-                                    )
-
-                                # Accumulate the tool call data
-                                if tc_chunk.id:
-                                    tool_calls[idx]["id"] = tc_chunk.id
-                                if tc_chunk.function:
-                                    if tc_chunk.function.name:
-                                        tool_calls[idx]["function"][
-                                            "name"
-                                        ] = tc_chunk.function.name
-                                    if tc_chunk.function.arguments:
-                                        tool_calls[idx]["function"][
-                                            "arguments"
-                                        ] += tc_chunk.function.arguments
-
-                                # Emit StreamToolInputStart only after we have the tool call ID
-                                if (
-                                    idx not in emitted_start_for_idx
-                                    and tool_calls[idx]["id"]
-                                    and tool_calls[idx]["function"]["name"]
-                                ):
-                                    yield StreamToolInputStart(
-                                        toolCallId=tool_calls[idx]["id"],
-                                        toolName=tool_calls[idx]["function"]["name"],
-                                    )
-                                    emitted_start_for_idx.add(idx)
-                logger.info(f"Stream complete. Finish reason: {finish_reason}")
-
-                # Yield all accumulated tool calls after the stream is complete
-                # This ensures all tool call arguments have been fully received
-                for idx, tool_call in enumerate(tool_calls):
-                    try:
-                        async for tc in _yield_tool_call(tool_calls, idx, session):
-                            yield tc
-                    except (orjson.JSONDecodeError, KeyError, TypeError) as e:
-                        logger.error(
-                            f"Failed to parse tool call {idx}: {e}",
-                            exc_info=True,
-                            extra={"tool_call": tool_call},
-                        )
-                        yield StreamError(
-                            errorText=f"Invalid tool call arguments for tool {tool_call.get('function', {}).get('name', 'unknown')}: {e}",
-                        )
-                        # Re-raise to trigger retry logic in the parent function
-                        raise
-
-                yield StreamFinish()
-                return
-            except Exception as e:
-                last_error = e
-                if _is_retryable_error(e) and retry_count < MAX_RETRIES:
-                    retry_count += 1
-                    # Calculate delay with exponential backoff
-                    delay = min(
-                        BASE_DELAY_SECONDS * (2 ** (retry_count - 1)),
-                        MAX_DELAY_SECONDS,
-                    )
-                    logger.warning(
-                        f"Retryable error in stream: {e!s}. "
-                        f"Retrying in {delay:.1f}s (attempt {retry_count}/{MAX_RETRIES})"
-                    )
-                    await asyncio.sleep(delay)
-                    continue  # Retry the stream
-                else:
-                    # Non-retryable error or max retries exceeded
-                    logger.error(
-                        f"Error in stream (not retrying): {e!s}",
-                        exc_info=True,
-                    )
-                    error_code = None
-                    error_text = str(e)
-                    if _is_region_blocked_error(e):
-                        error_code = "MODEL_NOT_AVAILABLE_REGION"
-                        error_text = (
-                            "This model is not available in your region. "
-                            "Please connect via VPN and try again."
-                        )
-                    error_response = StreamError(
-                        errorText=error_text,
-                        code=error_code,
-                    )
-                    yield error_response
-                    yield StreamFinish()
-                    return
-
-        # If we exit the retry loop without returning, it means we exhausted retries
-        if last_error:
-            logger.error(
-                f"Max retries ({MAX_RETRIES}) exceeded. Last error: {last_error!s}",
-                exc_info=True,
-            )
-            yield StreamError(errorText=f"Max retries exceeded: {last_error!s}")
-            yield StreamFinish()
-            return
-
-
-async def _yield_tool_call(
-    tool_calls: list[dict[str, Any]],
-    yield_idx: int,
-    session: ChatSession,
-) -> AsyncGenerator[StreamBaseResponse, None]:
-    """
-    Yield a tool call and its execution result.
-
-    Raises:
-        orjson.JSONDecodeError: If tool call arguments cannot be parsed as JSON
-        KeyError: If expected tool call fields are missing
-        TypeError: If tool call structure is invalid
-    """
-    tool_name = tool_calls[yield_idx]["function"]["name"]
-    tool_call_id = tool_calls[yield_idx]["id"]
-    logger.info(f"Yielding tool call: {tool_calls[yield_idx]}")
-
-    # Parse tool call arguments - handle empty arguments gracefully
-    raw_arguments = tool_calls[yield_idx]["function"]["arguments"]
-    if raw_arguments:
-        arguments = orjson.loads(raw_arguments)
-    else:
-        arguments = {}
-
-    yield StreamToolInputAvailable(
-        toolCallId=tool_call_id,
-        toolName=tool_name,
-        input=arguments,
-    )
-
-    tool_execution_response: StreamToolOutputAvailable = await execute_tool(
-        tool_name=tool_name,
-        parameters=arguments,
-        tool_call_id=tool_call_id,
-        user_id=session.user_id,
-        session=session,
-    )
-
-    yield tool_execution_response

autogpt_platform/backend/backend/api/features/chat/tools/agent_generator/__init__.py

@@ -1,29 +0,0 @@
-"""Agent generator package - Creates agents from natural language."""
-
-from .core import (
-    apply_agent_patch,
-    decompose_goal,
-    generate_agent,
-    generate_agent_patch,
-    get_agent_as_json,
-    save_agent_to_library,
-)
-from .fixer import apply_all_fixes
-from .utils import get_blocks_info
-from .validator import validate_agent
-
-__all__ = [
-    # Core functions
-    "decompose_goal",
-    "generate_agent",
-    "generate_agent_patch",
-    "apply_agent_patch",
-    "save_agent_to_library",
-    "get_agent_as_json",
-    # Fixer
-    "apply_all_fixes",
-    # Validator
-    "validate_agent",
-    # Utils
-    "get_blocks_info",
-]

autogpt_platform/backend/backend/api/features/chat/tools/agent_generator/client.py

@@ -1,25 +0,0 @@
-"""OpenRouter client configuration for agent generation."""
-
-import os
-
-from openai import AsyncOpenAI
-
-# Configuration - use OPEN_ROUTER_API_KEY for consistency with chat/config.py
-OPENROUTER_API_KEY = os.getenv("OPEN_ROUTER_API_KEY")
-AGENT_GENERATOR_MODEL = os.getenv("AGENT_GENERATOR_MODEL", "anthropic/claude-opus-4.5")
-
-# OpenRouter client (OpenAI-compatible API)
-_client: AsyncOpenAI | None = None
-
-
-def get_client() -> AsyncOpenAI:
-    """Get or create the OpenRouter client."""
-    global _client
-    if _client is None:
-        if not OPENROUTER_API_KEY:
-            raise ValueError("OPENROUTER_API_KEY environment variable is required")
-        _client = AsyncOpenAI(
-            base_url="https://openrouter.ai/api/v1",
-            api_key=OPENROUTER_API_KEY,
-        )
-    return _client

autogpt_platform/backend/backend/api/features/chat/tools/agent_generator/core.py

@@ -1,391 +0,0 @@
-"""Core agent generation functions."""
-
-import copy
-import json
-import logging
-import uuid
-from typing import Any
-
-from backend.api.features.library import db as library_db
-from backend.data.graph import Graph, Link, Node, create_graph
-
-from .client import AGENT_GENERATOR_MODEL, get_client
-from .prompts import DECOMPOSITION_PROMPT, GENERATION_PROMPT, PATCH_PROMPT
-from .utils import get_block_summaries, parse_json_from_llm
-
-logger = logging.getLogger(__name__)
-
-
-async def decompose_goal(description: str, context: str = "") -> dict[str, Any] | None:
-    """Break down a goal into steps or return clarifying questions.
-
-    Args:
-        description: Natural language goal description
-        context: Additional context (e.g., answers to previous questions)
-
-    Returns:
-        Dict with either:
-        - {"type": "clarifying_questions", "questions": [...]}
-        - {"type": "instructions", "steps": [...]}
-        Or None on error
-    """
-    client = get_client()
-    prompt = DECOMPOSITION_PROMPT.format(block_summaries=get_block_summaries())
-
-    full_description = description
-    if context:
-        full_description = f"{description}\n\nAdditional context:\n{context}"
-
-    try:
-        response = await client.chat.completions.create(
-            model=AGENT_GENERATOR_MODEL,
-            messages=[
-                {"role": "system", "content": prompt},
-                {"role": "user", "content": full_description},
-            ],
-            temperature=0,
-        )
-
-        content = response.choices[0].message.content
-        if content is None:
-            logger.error("LLM returned empty content for decomposition")
-            return None
-
-        result = parse_json_from_llm(content)
-
-        if result is None:
-            logger.error(f"Failed to parse decomposition response: {content[:200]}")
-            return None
-
-        return result
-
-    except Exception as e:
-        logger.error(f"Error decomposing goal: {e}")
-        return None
-
-
-async def generate_agent(instructions: dict[str, Any]) -> dict[str, Any] | None:
-    """Generate agent JSON from instructions.
-
-    Args:
-        instructions: Structured instructions from decompose_goal
-
-    Returns:
-        Agent JSON dict or None on error
-    """
-    client = get_client()
-    prompt = GENERATION_PROMPT.format(block_summaries=get_block_summaries())
-
-    try:
-        response = await client.chat.completions.create(
-            model=AGENT_GENERATOR_MODEL,
-            messages=[
-                {"role": "system", "content": prompt},
-                {"role": "user", "content": json.dumps(instructions, indent=2)},
-            ],
-            temperature=0,
-        )
-
-        content = response.choices[0].message.content
-        if content is None:
-            logger.error("LLM returned empty content for agent generation")
-            return None
-
-        result = parse_json_from_llm(content)
-
-        if result is None:
-            logger.error(f"Failed to parse agent JSON: {content[:200]}")
-            return None
-
-        # Ensure required fields
-        if "id" not in result:
-            result["id"] = str(uuid.uuid4())
-        if "version" not in result:
-            result["version"] = 1
-        if "is_active" not in result:
-            result["is_active"] = True
-
-        return result
-
-    except Exception as e:
-        logger.error(f"Error generating agent: {e}")
-        return None
-
-
-def json_to_graph(agent_json: dict[str, Any]) -> Graph:
-    """Convert agent JSON dict to Graph model.
-
-    Args:
-        agent_json: Agent JSON with nodes and links
-
-    Returns:
-        Graph ready for saving
-    """
-    nodes = []
-    for n in agent_json.get("nodes", []):
-        node = Node(
-            id=n.get("id", str(uuid.uuid4())),
-            block_id=n["block_id"],
-            input_default=n.get("input_default", {}),
-            metadata=n.get("metadata", {}),
-        )
-        nodes.append(node)
-
-    links = []
-    for link_data in agent_json.get("links", []):
-        link = Link(
-            id=link_data.get("id", str(uuid.uuid4())),
-            source_id=link_data["source_id"],
-            sink_id=link_data["sink_id"],
-            source_name=link_data["source_name"],
-            sink_name=link_data["sink_name"],
-            is_static=link_data.get("is_static", False),
-        )
-        links.append(link)
-
-    return Graph(
-        id=agent_json.get("id", str(uuid.uuid4())),
-        version=agent_json.get("version", 1),
-        is_active=agent_json.get("is_active", True),
-        name=agent_json.get("name", "Generated Agent"),
-        description=agent_json.get("description", ""),
-        nodes=nodes,
-        links=links,
-    )
-
-
-def _reassign_node_ids(graph: Graph) -> None:
-    """Reassign all node and link IDs to new UUIDs.
-
-    This is needed when creating a new version to avoid unique constraint violations.
-    """
-    # Create mapping from old node IDs to new UUIDs
-    id_map = {node.id: str(uuid.uuid4()) for node in graph.nodes}
-
-    # Reassign node IDs
-    for node in graph.nodes:
-        node.id = id_map[node.id]
-
-    # Update link references to use new node IDs
-    for link in graph.links:
-        link.id = str(uuid.uuid4())  # Also give links new IDs
-        if link.source_id in id_map:
-            link.source_id = id_map[link.source_id]
-        if link.sink_id in id_map:
-            link.sink_id = id_map[link.sink_id]
-
-
-async def save_agent_to_library(
-    agent_json: dict[str, Any], user_id: str, is_update: bool = False
-) -> tuple[Graph, Any]:
-    """Save agent to database and user's library.
-
-    Args:
-        agent_json: Agent JSON dict
-        user_id: User ID
-        is_update: Whether this is an update to an existing agent
-
-    Returns:
-        Tuple of (created Graph, LibraryAgent)
-    """
-    from backend.data.graph import get_graph_all_versions
-
-    graph = json_to_graph(agent_json)
-
-    if is_update:
-        # For updates, keep the same graph ID but increment version
-        # and reassign node/link IDs to avoid conflicts
-        if graph.id:
-            existing_versions = await get_graph_all_versions(graph.id, user_id)
-            if existing_versions:
-                latest_version = max(v.version for v in existing_versions)
-                graph.version = latest_version + 1
-                # Reassign node IDs (but keep graph ID the same)
-                _reassign_node_ids(graph)
-                logger.info(f"Updating agent {graph.id} to version {graph.version}")
-    else:
-        # For new agents, always generate a fresh UUID to avoid collisions
-        graph.id = str(uuid.uuid4())
-        graph.version = 1
-        # Reassign all node IDs as well
-        _reassign_node_ids(graph)
-        logger.info(f"Creating new agent with ID {graph.id}")
-
-    # Save to database
-    created_graph = await create_graph(graph, user_id)
-
-    # Add to user's library (or update existing library agent)
-    library_agents = await library_db.create_library_agent(
-        graph=created_graph,
-        user_id=user_id,
-        sensitive_action_safe_mode=True,
-        create_library_agents_for_sub_graphs=False,
-    )
-
-    return created_graph, library_agents[0]
-
-
-async def get_agent_as_json(
-    graph_id: str, user_id: str | None
-) -> dict[str, Any] | None:
-    """Fetch an agent and convert to JSON format for editing.
-
-    Args:
-        graph_id: Graph ID or library agent ID
-        user_id: User ID
-
-    Returns:
-        Agent as JSON dict or None if not found
-    """
-    from backend.data.graph import get_graph
-
-    # Try to get the graph (version=None gets the active version)
-    graph = await get_graph(graph_id, version=None, user_id=user_id)
-    if not graph:
-        return None
-
-    # Convert to JSON format
-    nodes = []
-    for node in graph.nodes:
-        nodes.append(
-            {
-                "id": node.id,
-                "block_id": node.block_id,
-                "input_default": node.input_default,
-                "metadata": node.metadata,
-            }
-        )
-
-    links = []
-    for node in graph.nodes:
-        for link in node.output_links:
-            links.append(
-                {
-                    "id": link.id,
-                    "source_id": link.source_id,
-                    "sink_id": link.sink_id,
-                    "source_name": link.source_name,
-                    "sink_name": link.sink_name,
-                    "is_static": link.is_static,
-                }
-            )
-
-    return {
-        "id": graph.id,
-        "name": graph.name,
-        "description": graph.description,
-        "version": graph.version,
-        "is_active": graph.is_active,
-        "nodes": nodes,
-        "links": links,
-    }
-
-
-async def generate_agent_patch(
-    update_request: str, current_agent: dict[str, Any]
-) -> dict[str, Any] | None:
-    """Generate a patch to update an existing agent.
-
-    Args:
-        update_request: Natural language description of changes
-        current_agent: Current agent JSON
-
-    Returns:
-        Patch dict or clarifying questions, or None on error
-    """
-    client = get_client()
-    prompt = PATCH_PROMPT.format(
-        current_agent=json.dumps(current_agent, indent=2),
-        block_summaries=get_block_summaries(),
-    )
-
-    try:
-        response = await client.chat.completions.create(
-            model=AGENT_GENERATOR_MODEL,
-            messages=[
-                {"role": "system", "content": prompt},
-                {"role": "user", "content": update_request},
-            ],
-            temperature=0,
-        )
-
-        content = response.choices[0].message.content
-        if content is None:
-            logger.error("LLM returned empty content for patch generation")
-            return None
-
-        return parse_json_from_llm(content)
-
-    except Exception as e:
-        logger.error(f"Error generating patch: {e}")
-        return None
-
-
-def apply_agent_patch(
-    current_agent: dict[str, Any], patch: dict[str, Any]
-) -> dict[str, Any]:
-    """Apply a patch to an existing agent.
-
-    Args:
-        current_agent: Current agent JSON
-        patch: Patch dict with operations
-
-    Returns:
-        Updated agent JSON
-    """
-    agent = copy.deepcopy(current_agent)
-    patches = patch.get("patches", [])
-
-    for p in patches:
-        patch_type = p.get("type")
-
-        if patch_type == "modify":
-            node_id = p.get("node_id")
-            changes = p.get("changes", {})
-
-            for node in agent.get("nodes", []):
-                if node["id"] == node_id:
-                    _deep_update(node, changes)
-                    logger.debug(f"Modified node {node_id}")
-                    break
-
-        elif patch_type == "add":
-            new_nodes = p.get("new_nodes", [])
-            new_links = p.get("new_links", [])
-
-            agent["nodes"] = agent.get("nodes", []) + new_nodes
-            agent["links"] = agent.get("links", []) + new_links
-            logger.debug(f"Added {len(new_nodes)} nodes, {len(new_links)} links")
-
-        elif patch_type == "remove":
-            node_ids_to_remove = set(p.get("node_ids", []))
-            link_ids_to_remove = set(p.get("link_ids", []))
-
-            # Remove nodes
-            agent["nodes"] = [
-                n for n in agent.get("nodes", []) if n["id"] not in node_ids_to_remove
-            ]
-
-            # Remove links (both explicit and those referencing removed nodes)
-            agent["links"] = [
-                link
-                for link in agent.get("links", [])
-                if link["id"] not in link_ids_to_remove
-                and link["source_id"] not in node_ids_to_remove
-                and link["sink_id"] not in node_ids_to_remove
-            ]
-
-            logger.debug(
-                f"Removed {len(node_ids_to_remove)} nodes, {len(link_ids_to_remove)} links"
-            )
-
-    return agent
-
-
-def _deep_update(target: dict, source: dict) -> None:
-    """Recursively update a dict with another dict."""
-    for key, value in source.items():
-        if key in target and isinstance(target[key], dict) and isinstance(value, dict):
-            _deep_update(target[key], value)
-        else:
-            target[key] = value

autogpt_platform/backend/backend/api/features/chat/tools/agent_generator/fixer.py

@@ -1,606 +0,0 @@
-"""Agent fixer - Fixes common LLM generation errors."""
-
-import logging
-import re
-import uuid
-from typing import Any
-
-from .utils import (
-    ADDTODICTIONARY_BLOCK_ID,
-    ADDTOLIST_BLOCK_ID,
-    CODE_EXECUTION_BLOCK_ID,
-    CONDITION_BLOCK_ID,
-    CREATEDICT_BLOCK_ID,
-    CREATELIST_BLOCK_ID,
-    DATA_SAMPLING_BLOCK_ID,
-    DOUBLE_CURLY_BRACES_BLOCK_IDS,
-    GET_CURRENT_DATE_BLOCK_ID,
-    STORE_VALUE_BLOCK_ID,
-    UNIVERSAL_TYPE_CONVERTER_BLOCK_ID,
-    get_blocks_info,
-    is_valid_uuid,
-)
-
-logger = logging.getLogger(__name__)
-
-
-def fix_agent_ids(agent: dict[str, Any]) -> dict[str, Any]:
-    """Fix invalid UUIDs in agent and link IDs."""
-    # Fix agent ID
-    if not is_valid_uuid(agent.get("id", "")):
-        agent["id"] = str(uuid.uuid4())
-        logger.debug(f"Fixed agent ID: {agent['id']}")
-
-    # Fix node IDs
-    id_mapping = {}  # Old ID -> New ID
-    for node in agent.get("nodes", []):
-        if not is_valid_uuid(node.get("id", "")):
-            old_id = node.get("id", "")
-            new_id = str(uuid.uuid4())
-            id_mapping[old_id] = new_id
-            node["id"] = new_id
-            logger.debug(f"Fixed node ID: {old_id} -> {new_id}")
-
-    # Fix link IDs and update references
-    for link in agent.get("links", []):
-        if not is_valid_uuid(link.get("id", "")):
-            link["id"] = str(uuid.uuid4())
-            logger.debug(f"Fixed link ID: {link['id']}")
-
-        # Update source/sink IDs if they were remapped
-        if link.get("source_id") in id_mapping:
-            link["source_id"] = id_mapping[link["source_id"]]
-        if link.get("sink_id") in id_mapping:
-            link["sink_id"] = id_mapping[link["sink_id"]]
-
-    return agent
-
-
-def fix_double_curly_braces(agent: dict[str, Any]) -> dict[str, Any]:
-    """Fix single curly braces to double in template blocks."""
-    for node in agent.get("nodes", []):
-        if node.get("block_id") not in DOUBLE_CURLY_BRACES_BLOCK_IDS:
-            continue
-
-        input_data = node.get("input_default", {})
-        for key in ("prompt", "format"):
-            if key in input_data and isinstance(input_data[key], str):
-                original = input_data[key]
-                # Fix simple variable references: {var} -> {{var}}
-                fixed = re.sub(
-                    r"(?<!\{)\{([a-zA-Z_][a-zA-Z0-9_]*)\}(?!\})",
-                    r"{{\1}}",
-                    original,
-                )
-                if fixed != original:
-                    input_data[key] = fixed
-                    logger.debug(f"Fixed curly braces in {key}")
-
-    return agent
-
-
-def fix_storevalue_before_condition(agent: dict[str, Any]) -> dict[str, Any]:
-    """Add StoreValueBlock before ConditionBlock if needed for value2."""
-    nodes = agent.get("nodes", [])
-    links = agent.get("links", [])
-
-    # Find all ConditionBlock nodes
-    condition_node_ids = {
-        node["id"] for node in nodes if node.get("block_id") == CONDITION_BLOCK_ID
-    }
-
-    if not condition_node_ids:
-        return agent
-
-    new_nodes = []
-    new_links = []
-    processed_conditions = set()
-
-    for link in links:
-        sink_id = link.get("sink_id")
-        sink_name = link.get("sink_name")
-
-        # Check if this link goes to a ConditionBlock's value2
-        if sink_id in condition_node_ids and sink_name == "value2":
-            source_node = next(
-                (n for n in nodes if n["id"] == link.get("source_id")), None
-            )
-
-            # Skip if source is already a StoreValueBlock
-            if source_node and source_node.get("block_id") == STORE_VALUE_BLOCK_ID:
-                continue
-
-            # Skip if we already processed this condition
-            if sink_id in processed_conditions:
-                continue
-
-            processed_conditions.add(sink_id)
-
-            # Create StoreValueBlock
-            store_node_id = str(uuid.uuid4())
-            store_node = {
-                "id": store_node_id,
-                "block_id": STORE_VALUE_BLOCK_ID,
-                "input_default": {"data": None},
-                "metadata": {"position": {"x": 0, "y": -100}},
-            }
-            new_nodes.append(store_node)
-
-            # Create link: original source -> StoreValueBlock
-            new_links.append(
-                {
-                    "id": str(uuid.uuid4()),
-                    "source_id": link["source_id"],
-                    "source_name": link["source_name"],
-                    "sink_id": store_node_id,
-                    "sink_name": "input",
-                    "is_static": False,
-                }
-            )
-
-            # Update original link: StoreValueBlock -> ConditionBlock
-            link["source_id"] = store_node_id
-            link["source_name"] = "output"
-
-            logger.debug(f"Added StoreValueBlock before ConditionBlock {sink_id}")
-
-    if new_nodes:
-        agent["nodes"] = nodes + new_nodes
-
-    return agent
-
-
-def fix_addtolist_blocks(agent: dict[str, Any]) -> dict[str, Any]:
-    """Fix AddToList blocks by adding prerequisite empty AddToList block.
-
-    When an AddToList block is found:
-    1. Checks if there's a CreateListBlock before it
-    2. Removes CreateListBlock if linked directly to AddToList
-    3. Adds an empty AddToList block before the original
-    4. Ensures the original has a self-referencing link
-    """
-    nodes = agent.get("nodes", [])
-    links = agent.get("links", [])
-    new_nodes = []
-    original_addtolist_ids = set()
-    nodes_to_remove = set()
-    links_to_remove = []
-
-    # First pass: identify CreateListBlock nodes to remove
-    for link in links:
-        source_node = next(
-            (n for n in nodes if n.get("id") == link.get("source_id")), None
-        )
-        sink_node = next((n for n in nodes if n.get("id") == link.get("sink_id")), None)
-
-        if (
-            source_node
-            and sink_node
-            and source_node.get("block_id") == CREATELIST_BLOCK_ID
-            and sink_node.get("block_id") == ADDTOLIST_BLOCK_ID
-        ):
-            nodes_to_remove.add(source_node.get("id"))
-            links_to_remove.append(link)
-            logger.debug(f"Removing CreateListBlock {source_node.get('id')}")
-
-    # Second pass: process AddToList blocks
-    filtered_nodes = []
-    for node in nodes:
-        if node.get("id") in nodes_to_remove:
-            continue
-
-        if node.get("block_id") == ADDTOLIST_BLOCK_ID:
-            original_addtolist_ids.add(node.get("id"))
-            node_id = node.get("id")
-            pos = node.get("metadata", {}).get("position", {"x": 0, "y": 0})
-
-            # Check if already has prerequisite
-            has_prereq = any(
-                link.get("sink_id") == node_id
-                and link.get("sink_name") == "list"
-                and link.get("source_name") == "updated_list"
-                for link in links
-            )
-
-            if not has_prereq:
-                # Remove links to "list" input (except self-reference)
-                for link in links:
-                    if (
-                        link.get("sink_id") == node_id
-                        and link.get("sink_name") == "list"
-                        and link.get("source_id") != node_id
-                        and link not in links_to_remove
-                    ):
-                        links_to_remove.append(link)
-
-                # Create prerequisite AddToList block
-                prereq_id = str(uuid.uuid4())
-                prereq_node = {
-                    "id": prereq_id,
-                    "block_id": ADDTOLIST_BLOCK_ID,
-                    "input_default": {"list": [], "entry": None, "entries": []},
-                    "metadata": {
-                        "position": {"x": pos.get("x", 0) - 800, "y": pos.get("y", 0)}
-                    },
-                }
-                new_nodes.append(prereq_node)
-
-                # Link prerequisite to original
-                links.append(
-                    {
-                        "id": str(uuid.uuid4()),
-                        "source_id": prereq_id,
-                        "source_name": "updated_list",
-                        "sink_id": node_id,
-                        "sink_name": "list",
-                        "is_static": False,
-                    }
-                )
-                logger.debug(f"Added prerequisite AddToList block for {node_id}")
-
-        filtered_nodes.append(node)
-
-    # Remove marked links
-    filtered_links = [link for link in links if link not in links_to_remove]
-
-    # Add self-referencing links for original AddToList blocks
-    for node in filtered_nodes + new_nodes:
-        if (
-            node.get("block_id") == ADDTOLIST_BLOCK_ID
-            and node.get("id") in original_addtolist_ids
-        ):
-            node_id = node.get("id")
-            has_self_ref = any(
-                link["source_id"] == node_id
-                and link["sink_id"] == node_id
-                and link["source_name"] == "updated_list"
-                and link["sink_name"] == "list"
-                for link in filtered_links
-            )
-            if not has_self_ref:
-                filtered_links.append(
-                    {
-                        "id": str(uuid.uuid4()),
-                        "source_id": node_id,
-                        "source_name": "updated_list",
-                        "sink_id": node_id,
-                        "sink_name": "list",
-                        "is_static": False,
-                    }
-                )
-                logger.debug(f"Added self-reference for AddToList {node_id}")
-
-    agent["nodes"] = filtered_nodes + new_nodes
-    agent["links"] = filtered_links
-    return agent
-
-
-def fix_addtodictionary_blocks(agent: dict[str, Any]) -> dict[str, Any]:
-    """Fix AddToDictionary blocks by removing empty CreateDictionary nodes."""
-    nodes = agent.get("nodes", [])
-    links = agent.get("links", [])
-    nodes_to_remove = set()
-    links_to_remove = []
-
-    for link in links:
-        source_node = next(
-            (n for n in nodes if n.get("id") == link.get("source_id")), None
-        )
-        sink_node = next((n for n in nodes if n.get("id") == link.get("sink_id")), None)
-
-        if (
-            source_node
-            and sink_node
-            and source_node.get("block_id") == CREATEDICT_BLOCK_ID
-            and sink_node.get("block_id") == ADDTODICTIONARY_BLOCK_ID
-        ):
-            nodes_to_remove.add(source_node.get("id"))
-            links_to_remove.append(link)
-            logger.debug(f"Removing CreateDictionary {source_node.get('id')}")
-
-    agent["nodes"] = [n for n in nodes if n.get("id") not in nodes_to_remove]
-    agent["links"] = [link for link in links if link not in links_to_remove]
-    return agent
-
-
-def fix_code_execution_output(agent: dict[str, Any]) -> dict[str, Any]:
-    """Fix CodeExecutionBlock output: change 'response' to 'stdout_logs'."""
-    nodes = agent.get("nodes", [])
-    links = agent.get("links", [])
-
-    for link in links:
-        source_node = next(
-            (n for n in nodes if n.get("id") == link.get("source_id")), None
-        )
-        if (
-            source_node
-            and source_node.get("block_id") == CODE_EXECUTION_BLOCK_ID
-            and link.get("source_name") == "response"
-        ):
-            link["source_name"] = "stdout_logs"
-            logger.debug("Fixed CodeExecutionBlock output: response -> stdout_logs")
-
-    return agent
-
-
-def fix_data_sampling_sample_size(agent: dict[str, Any]) -> dict[str, Any]:
-    """Fix DataSamplingBlock by setting sample_size to 1 as default."""
-    nodes = agent.get("nodes", [])
-    links = agent.get("links", [])
-    links_to_remove = []
-
-    for node in nodes:
-        if node.get("block_id") == DATA_SAMPLING_BLOCK_ID:
-            node_id = node.get("id")
-            input_default = node.get("input_default", {})
-
-            # Remove links to sample_size
-            for link in links:
-                if (
-                    link.get("sink_id") == node_id
-                    and link.get("sink_name") == "sample_size"
-                ):
-                    links_to_remove.append(link)
-
-            # Set default
-            input_default["sample_size"] = 1
-            node["input_default"] = input_default
-            logger.debug(f"Fixed DataSamplingBlock {node_id} sample_size to 1")
-
-    if links_to_remove:
-        agent["links"] = [link for link in links if link not in links_to_remove]
-
-    return agent
-
-
-def fix_node_x_coordinates(agent: dict[str, Any]) -> dict[str, Any]:
-    """Fix node x-coordinates to ensure 800+ unit spacing between linked nodes."""
-    nodes = agent.get("nodes", [])
-    links = agent.get("links", [])
-    node_lookup = {n.get("id"): n for n in nodes}
-
-    for link in links:
-        source_id = link.get("source_id")
-        sink_id = link.get("sink_id")
-
-        source_node = node_lookup.get(source_id)
-        sink_node = node_lookup.get(sink_id)
-
-        if not source_node or not sink_node:
-            continue
-
-        source_pos = source_node.get("metadata", {}).get("position", {})
-        sink_pos = sink_node.get("metadata", {}).get("position", {})
-
-        source_x = source_pos.get("x", 0)
-        sink_x = sink_pos.get("x", 0)
-
-        if abs(sink_x - source_x) < 800:
-            new_x = source_x + 800
-            if "metadata" not in sink_node:
-                sink_node["metadata"] = {}
-            if "position" not in sink_node["metadata"]:
-                sink_node["metadata"]["position"] = {}
-            sink_node["metadata"]["position"]["x"] = new_x
-            logger.debug(f"Fixed node {sink_id} x: {sink_x} -> {new_x}")
-
-    return agent
-
-
-def fix_getcurrentdate_offset(agent: dict[str, Any]) -> dict[str, Any]:
-    """Fix GetCurrentDateBlock offset to ensure it's positive."""
-    for node in agent.get("nodes", []):
-        if node.get("block_id") == GET_CURRENT_DATE_BLOCK_ID:
-            input_default = node.get("input_default", {})
-            if "offset" in input_default:
-                offset = input_default["offset"]
-                if isinstance(offset, (int, float)) and offset < 0:
-                    input_default["offset"] = abs(offset)
-                    logger.debug(f"Fixed offset: {offset} -> {abs(offset)}")
-
-    return agent
-
-
-def fix_ai_model_parameter(
-    agent: dict[str, Any],
-    blocks_info: list[dict[str, Any]],
-    default_model: str = "gpt-4o",
-) -> dict[str, Any]:
-    """Add default model parameter to AI blocks if missing."""
-    block_map = {b.get("id"): b for b in blocks_info}
-
-    for node in agent.get("nodes", []):
-        block_id = node.get("block_id")
-        block = block_map.get(block_id)
-
-        if not block:
-            continue
-
-        # Check if block has AI category
-        categories = block.get("categories", [])
-        is_ai_block = any(
-            cat.get("category") == "AI" for cat in categories if isinstance(cat, dict)
-        )
-
-        if is_ai_block:
-            input_default = node.get("input_default", {})
-            if "model" not in input_default:
-                input_default["model"] = default_model
-                node["input_default"] = input_default
-                logger.debug(
-                    f"Added model '{default_model}' to AI block {node.get('id')}"
-                )
-
-    return agent
-
-
-def fix_link_static_properties(
-    agent: dict[str, Any], blocks_info: list[dict[str, Any]]
-) -> dict[str, Any]:
-    """Fix is_static property based on source block's staticOutput."""
-    block_map = {b.get("id"): b for b in blocks_info}
-    node_lookup = {n.get("id"): n for n in agent.get("nodes", [])}
-
-    for link in agent.get("links", []):
-        source_node = node_lookup.get(link.get("source_id"))
-        if not source_node:
-            continue
-
-        source_block = block_map.get(source_node.get("block_id"))
-        if not source_block:
-            continue
-
-        static_output = source_block.get("staticOutput", False)
-        if link.get("is_static") != static_output:
-            link["is_static"] = static_output
-            logger.debug(f"Fixed link {link.get('id')} is_static to {static_output}")
-
-    return agent
-
-
-def fix_data_type_mismatch(
-    agent: dict[str, Any], blocks_info: list[dict[str, Any]]
-) -> dict[str, Any]:
-    """Fix data type mismatches by inserting UniversalTypeConverterBlock."""
-    nodes = agent.get("nodes", [])
-    links = agent.get("links", [])
-    block_map = {b.get("id"): b for b in blocks_info}
-    node_lookup = {n.get("id"): n for n in nodes}
-
-    def get_property_type(schema: dict, name: str) -> str | None:
-        if "_#_" in name:
-            parent, child = name.split("_#_", 1)
-            parent_schema = schema.get(parent, {})
-            if "properties" in parent_schema:
-                return parent_schema["properties"].get(child, {}).get("type")
-            return None
-        return schema.get(name, {}).get("type")
-
-    def are_types_compatible(src: str, sink: str) -> bool:
-        if {src, sink} <= {"integer", "number"}:
-            return True
-        return src == sink
-
-    type_mapping = {
-        "string": "string",
-        "text": "string",
-        "integer": "number",
-        "number": "number",
-        "float": "number",
-        "boolean": "boolean",
-        "bool": "boolean",
-        "array": "list",
-        "list": "list",
-        "object": "dictionary",
-        "dict": "dictionary",
-        "dictionary": "dictionary",
-    }
-
-    new_links = []
-    nodes_to_add = []
-
-    for link in links:
-        source_node = node_lookup.get(link.get("source_id"))
-        sink_node = node_lookup.get(link.get("sink_id"))
-
-        if not source_node or not sink_node:
-            new_links.append(link)
-            continue
-
-        source_block = block_map.get(source_node.get("block_id"))
-        sink_block = block_map.get(sink_node.get("block_id"))
-
-        if not source_block or not sink_block:
-            new_links.append(link)
-            continue
-
-        source_outputs = source_block.get("outputSchema", {}).get("properties", {})
-        sink_inputs = sink_block.get("inputSchema", {}).get("properties", {})
-
-        source_type = get_property_type(source_outputs, link.get("source_name", ""))
-        sink_type = get_property_type(sink_inputs, link.get("sink_name", ""))
-
-        if (
-            source_type
-            and sink_type
-            and not are_types_compatible(source_type, sink_type)
-        ):
-            # Insert type converter
-            converter_id = str(uuid.uuid4())
-            target_type = type_mapping.get(sink_type, sink_type)
-
-            converter_node = {
-                "id": converter_id,
-                "block_id": UNIVERSAL_TYPE_CONVERTER_BLOCK_ID,
-                "input_default": {"type": target_type},
-                "metadata": {"position": {"x": 0, "y": 100}},
-            }
-            nodes_to_add.append(converter_node)
-
-            # source -> converter
-            new_links.append(
-                {
-                    "id": str(uuid.uuid4()),
-                    "source_id": link["source_id"],
-                    "source_name": link["source_name"],
-                    "sink_id": converter_id,
-                    "sink_name": "value",
-                    "is_static": False,
-                }
-            )
-
-            # converter -> sink
-            new_links.append(
-                {
-                    "id": str(uuid.uuid4()),
-                    "source_id": converter_id,
-                    "source_name": "value",
-                    "sink_id": link["sink_id"],
-                    "sink_name": link["sink_name"],
-                    "is_static": False,
-                }
-            )
-
-            logger.debug(f"Inserted type converter: {source_type} -> {target_type}")
-        else:
-            new_links.append(link)
-
-    if nodes_to_add:
-        agent["nodes"] = nodes + nodes_to_add
-        agent["links"] = new_links
-
-    return agent
-
-
-def apply_all_fixes(
-    agent: dict[str, Any], blocks_info: list[dict[str, Any]] | None = None
-) -> dict[str, Any]:
-    """Apply all fixes to an agent JSON.
-
-    Args:
-        agent: Agent JSON dict
-        blocks_info: Optional list of block info dicts for advanced fixes
-
-    Returns:
-        Fixed agent JSON
-    """
-    # Basic fixes (no block info needed)
-    agent = fix_agent_ids(agent)
-    agent = fix_double_curly_braces(agent)
-    agent = fix_storevalue_before_condition(agent)
-    agent = fix_addtolist_blocks(agent)
-    agent = fix_addtodictionary_blocks(agent)
-    agent = fix_code_execution_output(agent)
-    agent = fix_data_sampling_sample_size(agent)
-    agent = fix_node_x_coordinates(agent)
-    agent = fix_getcurrentdate_offset(agent)
-
-    # Advanced fixes (require block info)
-    if blocks_info is None:
-        blocks_info = get_blocks_info()
-
-    agent = fix_ai_model_parameter(agent, blocks_info)
-    agent = fix_link_static_properties(agent, blocks_info)
-    agent = fix_data_type_mismatch(agent, blocks_info)
-
-    return agent

autogpt_platform/backend/backend/api/features/chat/tools/agent_generator/prompts.py

@@ -1,225 +0,0 @@
-"""Prompt templates for agent generation."""
-
-DECOMPOSITION_PROMPT = """
-You are an expert AutoGPT Workflow Decomposer. Your task is to analyze a user's high-level goal and break it down into a clear, step-by-step plan using the available blocks.
-
-Each step should represent a distinct, automatable action suitable for execution by an AI automation system.
-
----
-
-FIRST: Analyze the user's goal and determine:
-1) Design-time configuration (fixed settings that won't change per run)
-2) Runtime inputs (values the agent's end-user will provide each time it runs)
-
-For anything that can vary per run (email addresses, names, dates, search terms, etc.):
-- DO NOT ask for the actual value
-- Instead, define it as an Agent Input with a clear name, type, and description
-
-Only ask clarifying questions about design-time config that affects how you build the workflow:
-- Which external service to use (e.g., "Gmail vs Outlook", "Notion vs Google Docs")
-- Required formats or structures (e.g., "CSV, JSON, or PDF output?")
-- Business rules that must be hard-coded
-
-IMPORTANT CLARIFICATIONS POLICY:
-- Ask no more than five essential questions
-- Do not ask for concrete values that can be provided at runtime as Agent Inputs
-- Do not ask for API keys or credentials; the platform handles those directly
-- If there is enough information to infer reasonable defaults, prefer to propose defaults
-
----
-
-GUIDELINES:
-1. List each step as a numbered item
-2. Describe the action clearly and specify inputs/outputs
-3. Ensure steps are in logical, sequential order
-4. Mention block names naturally (e.g., "Use GetWeatherByLocationBlock to...")
-5. Help the user reach their goal efficiently
-
----
-
-RULES:
-1. OUTPUT FORMAT: Only output either clarifying questions OR step-by-step instructions, not both
-2. USE ONLY THE BLOCKS PROVIDED
-3. ALL required_input fields must be provided
-4. Data types of linked properties must match
-5. Write expert-level prompts for AI-related blocks
-
----
-
-CRITICAL BLOCK RESTRICTIONS:
-1. AddToListBlock: Outputs updated list EVERY addition, not after all additions
-2. SendEmailBlock: Draft the email for user review; set SMTP config based on email type
-3. ConditionBlock: value2 is reference, value1 is contrast
-4. CodeExecutionBlock: DO NOT USE - use AI blocks instead
-5. ReadCsvBlock: Only use the 'rows' output, not 'row'
-
----
-
-OUTPUT FORMAT:
-
-If more information is needed:
-```json
-{{
-  "type": "clarifying_questions",
-  "questions": [
-    {{
-      "question": "Which email provider should be used? (Gmail, Outlook, custom SMTP)",
-      "keyword": "email_provider",
-      "example": "Gmail"
-    }}
-  ]
-}}
-```
-
-If ready to proceed:
-```json
-{{
-  "type": "instructions",
-  "steps": [
-    {{
-      "step_number": 1,
-      "block_name": "AgentShortTextInputBlock",
-      "description": "Get the URL of the content to analyze.",
-      "inputs": [{{"name": "name", "value": "URL"}}],
-      "outputs": [{{"name": "result", "description": "The URL entered by user"}}]
-    }}
-  ]
-}}
-```
-
----
-
-AVAILABLE BLOCKS:
-{block_summaries}
-"""
-
-GENERATION_PROMPT = """
-You are an expert AI workflow builder. Generate a valid agent JSON from the given instructions.
-
----
-
-NODES:
-Each node must include:
-- `id`: Unique UUID v4 (e.g. `a8f5b1e2-c3d4-4e5f-8a9b-0c1d2e3f4a5b`)
-- `block_id`: The block identifier (must match an Allowed Block)
-- `input_default`: Dict of inputs (can be empty if no static inputs needed)
-- `metadata`: Must contain:
-  - `position`: {{"x": number, "y": number}} - adjacent nodes should differ by 800+ in X
-  - `customized_name`: Clear name describing this block's purpose in the workflow
-
----
-
-LINKS:
-Each link connects a source node's output to a sink node's input:
-- `id`: MUST be UUID v4 (NOT "link-1", "link-2", etc.)
-- `source_id`: ID of the source node
-- `source_name`: Output field name from the source block
-- `sink_id`: ID of the sink node
-- `sink_name`: Input field name on the sink block
-- `is_static`: true only if source block has static_output: true
-
-CRITICAL: All IDs must be valid UUID v4 format!
-
----
-
-AGENT (GRAPH):
-Wrap nodes and links in:
-- `id`: UUID of the agent
-- `name`: Short, generic name (avoid specific company names, URLs)
-- `description`: Short, generic description
-- `nodes`: List of all nodes
-- `links`: List of all links
-- `version`: 1
-- `is_active`: true
-
----
-
-TIPS:
-- All required_input fields must be provided via input_default or a valid link
-- Ensure consistent source_id and sink_id references
-- Avoid dangling links
-- Input/output pins must match block schemas
-- Do not invent unknown block_ids
-
----
-
-ALLOWED BLOCKS:
-{block_summaries}
-
----
-
-Generate the complete agent JSON. Output ONLY valid JSON, no explanation.
-"""
-
-PATCH_PROMPT = """
-You are an expert at modifying AutoGPT agent workflows. Given the current agent and a modification request, generate a JSON patch to update the agent.
-
-CURRENT AGENT:
-{current_agent}
-
-AVAILABLE BLOCKS:
-{block_summaries}
-
----
-
-PATCH FORMAT:
-Return a JSON object with the following structure:
-
-```json
-{{
-  "type": "patch",
-  "intent": "Brief description of what the patch does",
-  "patches": [
-    {{
-      "type": "modify",
-      "node_id": "uuid-of-node-to-modify",
-      "changes": {{
-        "input_default": {{"field": "new_value"}},
-        "metadata": {{"customized_name": "New Name"}}
-      }}
-    }},
-    {{
-      "type": "add",
-      "new_nodes": [
-        {{
-          "id": "new-uuid",
-          "block_id": "block-uuid",
-          "input_default": {{}},
-          "metadata": {{"position": {{"x": 0, "y": 0}}, "customized_name": "Name"}}
-        }}
-      ],
-      "new_links": [
-        {{
-          "id": "link-uuid",
-          "source_id": "source-node-id",
-          "source_name": "output_field",
-          "sink_id": "sink-node-id",
-          "sink_name": "input_field"
-        }}
-      ]
-    }},
-    {{
-      "type": "remove",
-      "node_ids": ["uuid-of-node-to-remove"],
-      "link_ids": ["uuid-of-link-to-remove"]
-    }}
-  ]
-}}
-```
-
-If you need more information, return:
-```json
-{{
-  "type": "clarifying_questions",
-  "questions": [
-    {{
-      "question": "What specific change do you want?",
-      "keyword": "change_type",
-      "example": "Add error handling"
-    }}
-  ]
-}}
-```
-
-Generate the minimal patch needed. Output ONLY valid JSON.
-"""

autogpt_platform/backend/backend/api/features/chat/tools/agent_generator/utils.py

@@ -1,213 +0,0 @@
-"""Utilities for agent generation."""
-
-import json
-import re
-from typing import Any
-
-from backend.data.block import get_blocks
-
-# UUID validation regex
-UUID_REGEX = re.compile(
-    r"^[a-f0-9]{8}-[a-f0-9]{4}-4[a-f0-9]{3}-[89ab][a-f0-9]{3}-[a-f0-9]{12}$"
-)
-
-# Block IDs for various fixes
-STORE_VALUE_BLOCK_ID = "1ff065e9-88e8-4358-9d82-8dc91f622ba9"
-CONDITION_BLOCK_ID = "715696a0-e1da-45c8-b209-c2fa9c3b0be6"
-ADDTOLIST_BLOCK_ID = "aeb08fc1-2fc1-4141-bc8e-f758f183a822"
-ADDTODICTIONARY_BLOCK_ID = "31d1064e-7446-4693-a7d4-65e5ca1180d1"
-CREATELIST_BLOCK_ID = "a912d5c7-6e00-4542-b2a9-8034136930e4"
-CREATEDICT_BLOCK_ID = "b924ddf4-de4f-4b56-9a85-358930dcbc91"
-CODE_EXECUTION_BLOCK_ID = "0b02b072-abe7-11ef-8372-fb5d162dd712"
-DATA_SAMPLING_BLOCK_ID = "4a448883-71fa-49cf-91cf-70d793bd7d87"
-UNIVERSAL_TYPE_CONVERTER_BLOCK_ID = "95d1b990-ce13-4d88-9737-ba5c2070c97b"
-GET_CURRENT_DATE_BLOCK_ID = "b29c1b50-5d0e-4d9f-8f9d-1b0e6fcbf0b1"
-
-DOUBLE_CURLY_BRACES_BLOCK_IDS = [
-    "44f6c8ad-d75c-4ae1-8209-aad1c0326928",  # FillTextTemplateBlock
-    "6ab085e2-20b3-4055-bc3e-08036e01eca6",
-    "90f8c45e-e983-4644-aa0b-b4ebe2f531bc",
-    "363ae599-353e-4804-937e-b2ee3cef3da4",  # AgentOutputBlock
-    "3b191d9f-356f-482d-8238-ba04b6d18381",
-    "db7d8f02-2f44-4c55-ab7a-eae0941f0c30",
-    "3a7c4b8d-6e2f-4a5d-b9c1-f8d23c5a9b0e",
-    "ed1ae7a0-b770-4089-b520-1f0005fad19a",
-    "a892b8d9-3e4e-4e9c-9c1e-75f8efcf1bfa",
-    "b29c1b50-5d0e-4d9f-8f9d-1b0e6fcbf0b1",
-    "716a67b3-6760-42e7-86dc-18645c6e00fc",
-    "530cf046-2ce0-4854-ae2c-659db17c7a46",
-    "ed55ac19-356e-4243-a6cb-bc599e9b716f",
-    "1f292d4a-41a4-4977-9684-7c8d560b9f91",  # LLM blocks
-    "32a87eab-381e-4dd4-bdb8-4c47151be35a",
-]
-
-
-def is_valid_uuid(value: str) -> bool:
-    """Check if a string is a valid UUID v4."""
-    return isinstance(value, str) and UUID_REGEX.match(value) is not None
-
-
-def _compact_schema(schema: dict) -> dict[str, str]:
-    """Extract compact type info from a JSON schema properties dict.
-
-    Returns a dict of {field_name: type_string} for essential info only.
-    """
-    props = schema.get("properties", {})
-    result = {}
-
-    for name, prop in props.items():
-        # Skip internal/complex fields
-        if name.startswith("_"):
-            continue
-
-        # Get type string
-        type_str = prop.get("type", "any")
-
-        # Handle anyOf/oneOf (optional types)
-        if "anyOf" in prop:
-            types = [t.get("type", "?") for t in prop["anyOf"] if t.get("type")]
-            type_str = "|".join(types) if types else "any"
-        elif "allOf" in prop:
-            type_str = "object"
-
-        # Add array item type if present
-        if type_str == "array" and "items" in prop:
-            items = prop["items"]
-            if isinstance(items, dict):
-                item_type = items.get("type", "any")
-                type_str = f"array[{item_type}]"
-
-        result[name] = type_str
-
-    return result
-
-
-def get_block_summaries(include_schemas: bool = True) -> str:
-    """Generate compact block summaries for prompts.
-
-    Args:
-        include_schemas: Whether to include input/output type info
-
-    Returns:
-        Formatted string of block summaries (compact format)
-    """
-    blocks = get_blocks()
-    summaries = []
-
-    for block_id, block_cls in blocks.items():
-        block = block_cls()
-        name = block.name
-        desc = getattr(block, "description", "") or ""
-
-        # Truncate description
-        if len(desc) > 150:
-            desc = desc[:147] + "..."
-
-        if not include_schemas:
-            summaries.append(f"- {name} (id: {block_id}): {desc}")
-        else:
-            # Compact format with type info only
-            inputs = {}
-            outputs = {}
-            required = []
-
-            if hasattr(block, "input_schema"):
-                try:
-                    schema = block.input_schema.jsonschema()
-                    inputs = _compact_schema(schema)
-                    required = schema.get("required", [])
-                except Exception:
-                    pass
-
-            if hasattr(block, "output_schema"):
-                try:
-                    schema = block.output_schema.jsonschema()
-                    outputs = _compact_schema(schema)
-                except Exception:
-                    pass
-
-            # Build compact line format
-            # Format: NAME (id): desc | in: {field:type, ...} [required] | out: {field:type}
-            in_str = ", ".join(f"{k}:{v}" for k, v in inputs.items())
-            out_str = ", ".join(f"{k}:{v}" for k, v in outputs.items())
-            req_str = f" req=[{','.join(required)}]" if required else ""
-
-            static = " [static]" if getattr(block, "static_output", False) else ""
-
-            line = f"- {name} (id: {block_id}): {desc}"
-            if in_str:
-                line += f"\n  in: {{{in_str}}}{req_str}"
-            if out_str:
-                line += f"\n  out: {{{out_str}}}{static}"
-
-            summaries.append(line)
-
-    return "\n".join(summaries)
-
-
-def get_blocks_info() -> list[dict[str, Any]]:
-    """Get block information with schemas for validation and fixing."""
-    blocks = get_blocks()
-    blocks_info = []
-    for block_id, block_cls in blocks.items():
-        block = block_cls()
-        blocks_info.append(
-            {
-                "id": block_id,
-                "name": block.name,
-                "description": getattr(block, "description", ""),
-                "categories": getattr(block, "categories", []),
-                "staticOutput": getattr(block, "static_output", False),
-                "inputSchema": (
-                    block.input_schema.jsonschema()
-                    if hasattr(block, "input_schema")
-                    else {}
-                ),
-                "outputSchema": (
-                    block.output_schema.jsonschema()
-                    if hasattr(block, "output_schema")
-                    else {}
-                ),
-            }
-        )
-    return blocks_info
-
-
-def parse_json_from_llm(text: str) -> dict[str, Any] | None:
-    """Extract JSON from LLM response (handles markdown code blocks)."""
-    if not text:
-        return None
-
-    # Try fenced code block
-    match = re.search(r"```(?:json)?\s*([\s\S]*?)```", text, re.IGNORECASE)
-    if match:
-        try:
-            return json.loads(match.group(1).strip())
-        except json.JSONDecodeError:
-            pass
-
-    # Try raw text
-    try:
-        return json.loads(text.strip())
-    except json.JSONDecodeError:
-        pass
-
-    # Try finding {...} span
-    start = text.find("{")
-    end = text.rfind("}")
-    if start != -1 and end > start:
-        try:
-            return json.loads(text[start : end + 1])
-        except json.JSONDecodeError:
-            pass
-
-    # Try finding [...] span
-    start = text.find("[")
-    end = text.rfind("]")
-    if start != -1 and end > start:
-        try:
-            return json.loads(text[start : end + 1])
-        except json.JSONDecodeError:
-            pass
-
-    return None

autogpt_platform/backend/backend/api/features/chat/tools/agent_generator/validator.py

@@ -1,279 +0,0 @@
-"""Agent validator - Validates agent structure and connections."""
-
-import logging
-import re
-from typing import Any
-
-from .utils import get_blocks_info
-
-logger = logging.getLogger(__name__)
-
-
-class AgentValidator:
-    """Validator for AutoGPT agents with detailed error reporting."""
-
-    def __init__(self):
-        self.errors: list[str] = []
-
-    def add_error(self, error: str) -> None:
-        """Add an error message."""
-        self.errors.append(error)
-
-    def validate_block_existence(
-        self, agent: dict[str, Any], blocks_info: list[dict[str, Any]]
-    ) -> bool:
-        """Validate all block IDs exist in the blocks library."""
-        valid = True
-        valid_block_ids = {b.get("id") for b in blocks_info if b.get("id")}
-
-        for node in agent.get("nodes", []):
-            block_id = node.get("block_id")
-            node_id = node.get("id")
-
-            if not block_id:
-                self.add_error(f"Node '{node_id}' is missing 'block_id' field.")
-                valid = False
-                continue
-
-            if block_id not in valid_block_ids:
-                self.add_error(
-                    f"Node '{node_id}' references block_id '{block_id}' which does not exist."
-                )
-                valid = False
-
-        return valid
-
-    def validate_link_node_references(self, agent: dict[str, Any]) -> bool:
-        """Validate all node IDs referenced in links exist."""
-        valid = True
-        valid_node_ids = {n.get("id") for n in agent.get("nodes", []) if n.get("id")}
-
-        for link in agent.get("links", []):
-            link_id = link.get("id", "Unknown")
-            source_id = link.get("source_id")
-            sink_id = link.get("sink_id")
-
-            if not source_id:
-                self.add_error(f"Link '{link_id}' is missing 'source_id'.")
-                valid = False
-            elif source_id not in valid_node_ids:
-                self.add_error(
-                    f"Link '{link_id}' references non-existent source_id '{source_id}'."
-                )
-                valid = False
-
-            if not sink_id:
-                self.add_error(f"Link '{link_id}' is missing 'sink_id'.")
-                valid = False
-            elif sink_id not in valid_node_ids:
-                self.add_error(
-                    f"Link '{link_id}' references non-existent sink_id '{sink_id}'."
-                )
-                valid = False
-
-        return valid
-
-    def validate_required_inputs(
-        self, agent: dict[str, Any], blocks_info: list[dict[str, Any]]
-    ) -> bool:
-        """Validate required inputs are provided."""
-        valid = True
-        block_map = {b.get("id"): b for b in blocks_info}
-
-        for node in agent.get("nodes", []):
-            block_id = node.get("block_id")
-            block = block_map.get(block_id)
-
-            if not block:
-                continue
-
-            required_inputs = block.get("inputSchema", {}).get("required", [])
-            input_defaults = node.get("input_default", {})
-            node_id = node.get("id")
-
-            # Get linked inputs
-            linked_inputs = {
-                link["sink_name"]
-                for link in agent.get("links", [])
-                if link.get("sink_id") == node_id
-            }
-
-            for req_input in required_inputs:
-                if (
-                    req_input not in input_defaults
-                    and req_input not in linked_inputs
-                    and req_input != "credentials"
-                ):
-                    block_name = block.get("name", "Unknown Block")
-                    self.add_error(
-                        f"Node '{node_id}' ({block_name}) is missing required input '{req_input}'."
-                    )
-                    valid = False
-
-        return valid
-
-    def validate_data_type_compatibility(
-        self, agent: dict[str, Any], blocks_info: list[dict[str, Any]]
-    ) -> bool:
-        """Validate linked data types are compatible."""
-        valid = True
-        block_map = {b.get("id"): b for b in blocks_info}
-        node_lookup = {n.get("id"): n for n in agent.get("nodes", [])}
-
-        def get_type(schema: dict, name: str) -> str | None:
-            if "_#_" in name:
-                parent, child = name.split("_#_", 1)
-                parent_schema = schema.get(parent, {})
-                if "properties" in parent_schema:
-                    return parent_schema["properties"].get(child, {}).get("type")
-                return None
-            return schema.get(name, {}).get("type")
-
-        def are_compatible(src: str, sink: str) -> bool:
-            if {src, sink} <= {"integer", "number"}:
-                return True
-            return src == sink
-
-        for link in agent.get("links", []):
-            source_node = node_lookup.get(link.get("source_id"))
-            sink_node = node_lookup.get(link.get("sink_id"))
-
-            if not source_node or not sink_node:
-                continue
-
-            source_block = block_map.get(source_node.get("block_id"))
-            sink_block = block_map.get(sink_node.get("block_id"))
-
-            if not source_block or not sink_block:
-                continue
-
-            source_outputs = source_block.get("outputSchema", {}).get("properties", {})
-            sink_inputs = sink_block.get("inputSchema", {}).get("properties", {})
-
-            source_type = get_type(source_outputs, link.get("source_name", ""))
-            sink_type = get_type(sink_inputs, link.get("sink_name", ""))
-
-            if source_type and sink_type and not are_compatible(source_type, sink_type):
-                self.add_error(
-                    f"Type mismatch: {source_block.get('name')} output '{link['source_name']}' "
-                    f"({source_type}) -> {sink_block.get('name')} input '{link['sink_name']}' ({sink_type})."
-                )
-                valid = False
-
-        return valid
-
-    def validate_nested_sink_links(
-        self, agent: dict[str, Any], blocks_info: list[dict[str, Any]]
-    ) -> bool:
-        """Validate nested sink links (with _#_ notation)."""
-        valid = True
-        block_map = {b.get("id"): b for b in blocks_info}
-        node_lookup = {n.get("id"): n for n in agent.get("nodes", [])}
-
-        for link in agent.get("links", []):
-            sink_name = link.get("sink_name", "")
-
-            if "_#_" in sink_name:
-                parent, child = sink_name.split("_#_", 1)
-
-                sink_node = node_lookup.get(link.get("sink_id"))
-                if not sink_node:
-                    continue
-
-                block = block_map.get(sink_node.get("block_id"))
-                if not block:
-                    continue
-
-                input_props = block.get("inputSchema", {}).get("properties", {})
-                parent_schema = input_props.get(parent)
-
-                if not parent_schema:
-                    self.add_error(
-                        f"Invalid nested link '{sink_name}': parent '{parent}' not found."
-                    )
-                    valid = False
-                    continue
-
-                if not parent_schema.get("additionalProperties"):
-                    if not (
-                        isinstance(parent_schema, dict)
-                        and "properties" in parent_schema
-                        and child in parent_schema.get("properties", {})
-                    ):
-                        self.add_error(
-                            f"Invalid nested link '{sink_name}': child '{child}' not found in '{parent}'."
-                        )
-                        valid = False
-
-        return valid
-
-    def validate_prompt_spaces(self, agent: dict[str, Any]) -> bool:
-        """Validate prompts don't have spaces in template variables."""
-        valid = True
-
-        for node in agent.get("nodes", []):
-            input_default = node.get("input_default", {})
-            prompt = input_default.get("prompt", "")
-
-            if not isinstance(prompt, str):
-                continue
-
-            # Find {{...}} with spaces
-            matches = re.finditer(r"\{\{([^}]+)\}\}", prompt)
-            for match in matches:
-                content = match.group(1)
-                if " " in content:
-                    self.add_error(
-                        f"Node '{node.get('id')}' has spaces in template variable: "
-                        f"'{{{{{content}}}}}' should be '{{{{{content.replace(' ', '_')}}}}}'."
-                    )
-                    valid = False
-
-        return valid
-
-    def validate(
-        self, agent: dict[str, Any], blocks_info: list[dict[str, Any]] | None = None
-    ) -> tuple[bool, str | None]:
-        """Run all validations.
-
-        Returns:
-            Tuple of (is_valid, error_message)
-        """
-        self.errors = []
-
-        if blocks_info is None:
-            blocks_info = get_blocks_info()
-
-        checks = [
-            self.validate_block_existence(agent, blocks_info),
-            self.validate_link_node_references(agent),
-            self.validate_required_inputs(agent, blocks_info),
-            self.validate_data_type_compatibility(agent, blocks_info),
-            self.validate_nested_sink_links(agent, blocks_info),
-            self.validate_prompt_spaces(agent),
-        ]
-
-        all_passed = all(checks)
-
-        if all_passed:
-            logger.info("Agent validation successful")
-            return True, None
-
-        error_message = "Agent validation failed:\n"
-        for i, error in enumerate(self.errors, 1):
-            error_message += f"{i}. {error}\n"
-
-        logger.warning(f"Agent validation failed with {len(self.errors)} errors")
-        return False, error_message
-
-
-def validate_agent(
-    agent: dict[str, Any], blocks_info: list[dict[str, Any]] | None = None
-) -> tuple[bool, str | None]:
-    """Convenience function to validate an agent.
-
-    Returns:
-        Tuple of (is_valid, error_message)
-    """
-    validator = AgentValidator()
-    return validator.validate(agent, blocks_info)

autogpt_platform/backend/backend/api/features/chat/tools/agent_search.py

@@ -1,151 +0,0 @@
-"""Shared agent search functionality for find_agent and find_library_agent tools."""
-
-import logging
-from typing import Literal
-
-from backend.api.features.library import db as library_db
-from backend.api.features.store import db as store_db
-from backend.util.exceptions import DatabaseError, NotFoundError
-
-from .models import (
-    AgentInfo,
-    AgentsFoundResponse,
-    ErrorResponse,
-    NoResultsResponse,
-    ToolResponseBase,
-)
-
-logger = logging.getLogger(__name__)
-
-SearchSource = Literal["marketplace", "library"]
-
-
-async def search_agents(
-    query: str,
-    source: SearchSource,
-    session_id: str | None,
-    user_id: str | None = None,
-) -> ToolResponseBase:
-    """
-    Search for agents in marketplace or user library.
-
-    Args:
-        query: Search query string
-        source: "marketplace" or "library"
-        session_id: Chat session ID
-        user_id: User ID (required for library search)
-
-    Returns:
-        AgentsFoundResponse, NoResultsResponse, or ErrorResponse
-    """
-    if not query:
-        return ErrorResponse(
-            message="Please provide a search query", session_id=session_id
-        )
-
-    if source == "library" and not user_id:
-        return ErrorResponse(
-            message="User authentication required to search library",
-            session_id=session_id,
-        )
-
-    agents: list[AgentInfo] = []
-    try:
-        if source == "marketplace":
-            logger.info(f"Searching marketplace for: {query}")
-            results = await store_db.get_store_agents(search_query=query, page_size=5)
-            for agent in results.agents:
-                agents.append(
-                    AgentInfo(
-                        id=f"{agent.creator}/{agent.slug}",
-                        name=agent.agent_name,
-                        description=agent.description or "",
-                        source="marketplace",
-                        in_library=False,
-                        creator=agent.creator,
-                        category="general",
-                        rating=agent.rating,
-                        runs=agent.runs,
-                        is_featured=False,
-                    )
-                )
-        else:  # library
-            logger.info(f"Searching user library for: {query}")
-            results = await library_db.list_library_agents(
-                user_id=user_id,  # type: ignore[arg-type]
-                search_term=query,
-                page_size=10,
-            )
-            for agent in results.agents:
-                agents.append(
-                    AgentInfo(
-                        id=agent.id,
-                        name=agent.name,
-                        description=agent.description or "",
-                        source="library",
-                        in_library=True,
-                        creator=agent.creator_name,
-                        status=agent.status.value,
-                        can_access_graph=agent.can_access_graph,
-                        has_external_trigger=agent.has_external_trigger,
-                        new_output=agent.new_output,
-                        graph_id=agent.graph_id,
-                    )
-                )
-        logger.info(f"Found {len(agents)} agents in {source}")
-    except NotFoundError:
-        pass
-    except DatabaseError as e:
-        logger.error(f"Error searching {source}: {e}", exc_info=True)
-        return ErrorResponse(
-            message=f"Failed to search {source}. Please try again.",
-            error=str(e),
-            session_id=session_id,
-        )
-
-    if not agents:
-        suggestions = (
-            [
-                "Try more general terms",
-                "Browse categories in the marketplace",
-                "Check spelling",
-            ]
-            if source == "marketplace"
-            else [
-                "Try different keywords",
-                "Use find_agent to search the marketplace",
-                "Check your library at /library",
-            ]
-        )
-        no_results_msg = (
-            f"No agents found matching '{query}'. Try different keywords or browse the marketplace."
-            if source == "marketplace"
-            else f"No agents matching '{query}' found in your library."
-        )
-        return NoResultsResponse(
-            message=no_results_msg, session_id=session_id, suggestions=suggestions
-        )
-
-    title = f"Found {len(agents)} agent{'s' if len(agents) != 1 else ''} "
-    title += (
-        f"for '{query}'"
-        if source == "marketplace"
-        else f"in your library for '{query}'"
-    )
-
-    message = (
-        "Now you have found some options for the user to choose from. "
-        "You can add a link to a recommended agent at: /marketplace/agent/agent_id "
-        "Please ask the user if they would like to use any of these agents."
-        if source == "marketplace"
-        else "Found agents in the user's library. You can provide a link to view an agent at: "
-        "/library/agents/{agent_id}. Use agent_output to get execution results, or run_agent to execute."
-    )
-
-    return AgentsFoundResponse(
-        message=message,
-        title=title,
-        agents=agents,
-        count=len(agents),
-        session_id=session_id,
-    )

autogpt_platform/backend/backend/api/features/chat/tools/find_block.py

@@ -1,194 +0,0 @@
-import logging
-from typing import Any
-
-from langfuse import observe
-from prisma.enums import ContentType
-
-from backend.api.features.chat.model import ChatSession
-from backend.api.features.chat.tools.base import BaseTool, ToolResponseBase
-from backend.api.features.chat.tools.models import (
-    BlockInfoSummary,
-    BlockInputFieldInfo,
-    BlockListResponse,
-    ErrorResponse,
-    NoResultsResponse,
-)
-from backend.api.features.store.hybrid_search import unified_hybrid_search
-from backend.data.block import get_block
-
-logger = logging.getLogger(__name__)
-
-
-class FindBlockTool(BaseTool):
-    """Tool for searching available blocks."""
-
-    @property
-    def name(self) -> str:
-        return "find_block"
-
-    @property
-    def description(self) -> str:
-        return (
-            "Search for available blocks by name or description. "
-            "Blocks are reusable components that perform specific tasks like "
-            "sending emails, making API calls, processing text, etc. "
-            "IMPORTANT: Use this tool FIRST to get the block's 'id' before calling run_block. "
-            "The response includes each block's id, required_inputs, and input_schema."
-        )
-
-    @property
-    def parameters(self) -> dict[str, Any]:
-        return {
-            "type": "object",
-            "properties": {
-                "query": {
-                    "type": "string",
-                    "description": (
-                        "Search query to find blocks by name or description. "
-                        "Use keywords like 'email', 'http', 'text', 'ai', etc."
-                    ),
-                },
-            },
-            "required": ["query"],
-        }
-
-    @property
-    def requires_auth(self) -> bool:
-        return True
-
-    @observe(as_type="tool", name="find_block")
-    async def _execute(
-        self,
-        user_id: str | None,
-        session: ChatSession,
-        **kwargs,
-    ) -> ToolResponseBase:
-        """Search for blocks matching the query.
-
-        Args:
-            user_id: User ID (required)
-            session: Chat session
-            query: Search query
-
-        Returns:
-            BlockListResponse: List of matching blocks
-            NoResultsResponse: No blocks found
-            ErrorResponse: Error message
-        """
-        query = kwargs.get("query", "").strip()
-        session_id = session.session_id
-
-        if not query:
-            return ErrorResponse(
-                message="Please provide a search query",
-                session_id=session_id,
-            )
-
-        try:
-            # Search for blocks using hybrid search
-            results, total = await unified_hybrid_search(
-                query=query,
-                content_types=[ContentType.BLOCK],
-                page=1,
-                page_size=10,
-            )
-
-            if not results:
-                return NoResultsResponse(
-                    message=f"No blocks found for '{query}'",
-                    suggestions=[
-                        "Try broader keywords like 'email', 'http', 'text', 'ai'",
-                        "Check spelling of technical terms",
-                    ],
-                    session_id=session_id,
-                )
-
-            # Enrich results with full block information
-            blocks: list[BlockInfoSummary] = []
-            for result in results:
-                block_id = result["content_id"]
-                block = get_block(block_id)
-
-                if block:
-                    # Get input/output schemas
-                    input_schema = {}
-                    output_schema = {}
-                    try:
-                        input_schema = block.input_schema.jsonschema()
-                    except Exception:
-                        pass
-                    try:
-                        output_schema = block.output_schema.jsonschema()
-                    except Exception:
-                        pass
-
-                    # Get categories from block instance
-                    categories = []
-                    if hasattr(block, "categories") and block.categories:
-                        categories = [cat.value for cat in block.categories]
-
-                    # Extract required inputs for easier use
-                    required_inputs: list[BlockInputFieldInfo] = []
-                    if input_schema:
-                        properties = input_schema.get("properties", {})
-                        required_fields = set(input_schema.get("required", []))
-                        # Get credential field names to exclude from required inputs
-                        credentials_fields = set(
-                            block.input_schema.get_credentials_fields().keys()
-                        )
-
-                        for field_name, field_schema in properties.items():
-                            # Skip credential fields - they're handled separately
-                            if field_name in credentials_fields:
-                                continue
-
-                            required_inputs.append(
-                                BlockInputFieldInfo(
-                                    name=field_name,
-                                    type=field_schema.get("type", "string"),
-                                    description=field_schema.get("description", ""),
-                                    required=field_name in required_fields,
-                                    default=field_schema.get("default"),
-                                )
-                            )
-
-                    blocks.append(
-                        BlockInfoSummary(
-                            id=block_id,
-                            name=block.name,
-                            description=block.description or "",
-                            categories=categories,
-                            input_schema=input_schema,
-                            output_schema=output_schema,
-                            required_inputs=required_inputs,
-                        )
-                    )
-
-            if not blocks:
-                return NoResultsResponse(
-                    message=f"No blocks found for '{query}'",
-                    suggestions=[
-                        "Try broader keywords like 'email', 'http', 'text', 'ai'",
-                    ],
-                    session_id=session_id,
-                )
-
-            return BlockListResponse(
-                message=(
-                    f"Found {len(blocks)} block(s) matching '{query}'. "
-                    "To execute a block, use run_block with the block's 'id' field "
-                    "and provide 'input_data' matching the block's input_schema."
-                ),
-                blocks=blocks,
-                count=len(blocks),
-                query=query,
-                session_id=session_id,
-            )
-
-        except Exception as e:
-            logger.error(f"Error searching blocks: {e}", exc_info=True)
-            return ErrorResponse(
-                message="Failed to search blocks",
-                error=str(e),
-                session_id=session_id,
-            )

autogpt_platform/backend/backend/api/features/executions/review/model.py

@@ -23,6 +23,7 @@ class PendingHumanReviewModel(BaseModel):
         id: Unique identifier for the review record
         user_id: ID of the user who must perform the review
         node_exec_id: ID of the node execution that created this review
+        node_id: ID of the node definition (for grouping reviews from same node)
         graph_exec_id: ID of the graph execution containing the node
         graph_id: ID of the graph template being executed
         graph_version: Version number of the graph template
@@ -37,6 +38,10 @@ class PendingHumanReviewModel(BaseModel):
     """
 
     node_exec_id: str = Field(description="Node execution ID (primary key)")
+    node_id: str = Field(
+        description="Node definition ID (for grouping)",
+        default="",  # Temporary default for test compatibility
+    )
     user_id: str = Field(description="User ID associated with the review")
     graph_exec_id: str = Field(description="Graph execution ID")
     graph_id: str = Field(description="Graph ID")
@@ -66,7 +71,9 @@ class PendingHumanReviewModel(BaseModel):
     )
 
     @classmethod
-    def from_db(cls, review: "PendingHumanReview") -> "PendingHumanReviewModel":
+    def from_db(
+        cls, review: "PendingHumanReview", node_id: str
+    ) -> "PendingHumanReviewModel":
         """
         Convert a database model to a response model.
 
@@ -74,9 +81,14 @@ class PendingHumanReviewModel(BaseModel):
         payload, instructions, and editable flag.
 
         Handles invalid data gracefully by using safe defaults.
+
+        Args:
+            review: Database review object
+            node_id: Node definition ID (fetched from NodeExecution)
         """
         return cls(
             node_exec_id=review.nodeExecId,
+            node_id=node_id,
             user_id=review.userId,
             graph_exec_id=review.graphExecId,
             graph_id=review.graphId,
@@ -107,6 +119,13 @@ class ReviewItem(BaseModel):
     reviewed_data: SafeJsonData | None = Field(
         None, description="Optional edited data (ignored if approved=False)"
     )
+    auto_approve_future: bool = Field(
+        default=False,
+        description=(
+            "If true and this review is approved, future executions of this same "
+            "block (node) will be automatically approved. This only affects approved reviews."
+        ),
+    )
 
     @field_validator("reviewed_data")
     @classmethod
@@ -174,6 +193,9 @@ class ReviewRequest(BaseModel):
     This request must include ALL pending reviews for a graph execution.
     Each review will be either approved (with optional data modifications)
     or rejected (data ignored). The execution will resume only after ALL reviews are processed.
+
+    Each review item can individually specify whether to auto-approve future executions
+    of the same block via the `auto_approve_future` field on ReviewItem.
     """
 
     reviews: List[ReviewItem] = Field(

autogpt_platform/backend/backend/api/features/executions/review/routes.py

@@ -1,17 +1,27 @@
+import asyncio
 import logging
-from typing import List
+from typing import Any, List
 
 import autogpt_libs.auth as autogpt_auth_lib
 from fastapi import APIRouter, HTTPException, Query, Security, status
 from prisma.enums import ReviewStatus
 
-from backend.data.execution import get_graph_execution_meta
+from backend.data.execution import (
+    ExecutionContext,
+    ExecutionStatus,
+    get_graph_execution_meta,
+)
+from backend.data.graph import get_graph_settings
 from backend.data.human_review import (
+    create_auto_approval_record,
     get_pending_reviews_for_execution,
     get_pending_reviews_for_user,
+    get_reviews_by_node_exec_ids,
     has_pending_reviews_for_graph_exec,
     process_all_reviews_for_execution,
 )
+from backend.data.model import USER_TIMEZONE_NOT_SET
+from backend.data.user import get_user_by_id
 from backend.executor.utils import add_graph_execution
 
 from .model import PendingHumanReviewModel, ReviewRequest, ReviewResponse
@@ -127,17 +137,70 @@ async def process_review_action(
             detail="At least one review must be provided",
         )
 
-    # Build review decisions map
+    # Batch fetch all requested reviews (regardless of status for idempotent handling)
+    reviews_map = await get_reviews_by_node_exec_ids(
+        list(all_request_node_ids), user_id
+    )
+
+    # Validate all reviews were found (must exist, any status is OK for now)
+    missing_ids = all_request_node_ids - set(reviews_map.keys())
+    if missing_ids:
+        raise HTTPException(
+            status_code=status.HTTP_404_NOT_FOUND,
+            detail=f"Review(s) not found: {', '.join(missing_ids)}",
+        )
+
+    # Validate all reviews belong to the same execution
+    graph_exec_ids = {review.graph_exec_id for review in reviews_map.values()}
+    if len(graph_exec_ids) > 1:
+        raise HTTPException(
+            status_code=status.HTTP_409_CONFLICT,
+            detail="All reviews in a single request must belong to the same execution.",
+        )
+
+    graph_exec_id = next(iter(graph_exec_ids))
+
+    # Validate execution status before processing reviews
+    graph_exec_meta = await get_graph_execution_meta(
+        user_id=user_id, execution_id=graph_exec_id
+    )
+
+    if not graph_exec_meta:
+        raise HTTPException(
+            status_code=status.HTTP_404_NOT_FOUND,
+            detail=f"Graph execution #{graph_exec_id} not found",
+        )
+
+    # Only allow processing reviews if execution is paused for review
+    # or incomplete (partial execution with some reviews already processed)
+    if graph_exec_meta.status not in (
+        ExecutionStatus.REVIEW,
+        ExecutionStatus.INCOMPLETE,
+    ):
+        raise HTTPException(
+            status_code=status.HTTP_409_CONFLICT,
+            detail=f"Cannot process reviews while execution status is {graph_exec_meta.status}. "
+            f"Reviews can only be processed when execution is paused (REVIEW status). "
+            f"Current status: {graph_exec_meta.status}",
+        )
+
+    # Build review decisions map and track which reviews requested auto-approval
+    # Auto-approved reviews use original data (no modifications allowed)
     review_decisions = {}
+    auto_approve_requests = {}  # Map node_exec_id -> auto_approve_future flag
+
     for review in request.reviews:
         review_status = (
             ReviewStatus.APPROVED if review.approved else ReviewStatus.REJECTED
         )
+        # If this review requested auto-approval, don't allow data modifications
+        reviewed_data = None if review.auto_approve_future else review.reviewed_data
         review_decisions[review.node_exec_id] = (
             review_status,
-            review.reviewed_data,
+            reviewed_data,
             review.message,
         )
+        auto_approve_requests[review.node_exec_id] = review.auto_approve_future
 
     # Process all reviews
     updated_reviews = await process_all_reviews_for_execution(
@@ -145,6 +208,87 @@ async def process_review_action(
         review_decisions=review_decisions,
     )
 
+    # Create auto-approval records for approved reviews that requested it
+    # Deduplicate by node_id to avoid race conditions when multiple reviews
+    # for the same node are processed in parallel
+    async def create_auto_approval_for_node(
+        node_id: str, review_result
+    ) -> tuple[str, bool]:
+        """
+        Create auto-approval record for a node.
+        Returns (node_id, success) tuple for tracking failures.
+        """
+        try:
+            await create_auto_approval_record(
+                user_id=user_id,
+                graph_exec_id=review_result.graph_exec_id,
+                graph_id=review_result.graph_id,
+                graph_version=review_result.graph_version,
+                node_id=node_id,
+                payload=review_result.payload,
+            )
+            return (node_id, True)
+        except Exception as e:
+            logger.error(
+                f"Failed to create auto-approval record for node {node_id}",
+                exc_info=e,
+            )
+            return (node_id, False)
+
+    # Collect node_exec_ids that need auto-approval
+    node_exec_ids_needing_auto_approval = [
+        node_exec_id
+        for node_exec_id, review_result in updated_reviews.items()
+        if review_result.status == ReviewStatus.APPROVED
+        and auto_approve_requests.get(node_exec_id, False)
+    ]
+
+    # Batch-fetch node executions to get node_ids
+    nodes_needing_auto_approval: dict[str, Any] = {}
+    if node_exec_ids_needing_auto_approval:
+        from backend.data.execution import get_node_executions
+
+        node_execs = await get_node_executions(
+            graph_exec_id=graph_exec_id, include_exec_data=False
+        )
+        node_exec_map = {node_exec.node_exec_id: node_exec for node_exec in node_execs}
+
+        for node_exec_id in node_exec_ids_needing_auto_approval:
+            node_exec = node_exec_map.get(node_exec_id)
+            if node_exec:
+                review_result = updated_reviews[node_exec_id]
+                # Use the first approved review for this node (deduplicate by node_id)
+                if node_exec.node_id not in nodes_needing_auto_approval:
+                    nodes_needing_auto_approval[node_exec.node_id] = review_result
+            else:
+                logger.error(
+                    f"Failed to create auto-approval record for {node_exec_id}: "
+                    f"Node execution not found. This may indicate a race condition "
+                    f"or data inconsistency."
+                )
+
+    # Execute all auto-approval creations in parallel (deduplicated by node_id)
+    auto_approval_results = await asyncio.gather(
+        *[
+            create_auto_approval_for_node(node_id, review_result)
+            for node_id, review_result in nodes_needing_auto_approval.items()
+        ],
+        return_exceptions=True,
+    )
+
+    # Count auto-approval failures
+    auto_approval_failed_count = 0
+    for result in auto_approval_results:
+        if isinstance(result, Exception):
+            # Unexpected exception during auto-approval creation
+            auto_approval_failed_count += 1
+            logger.error(
+                f"Unexpected exception during auto-approval creation: {result}"
+            )
+        elif isinstance(result, tuple) and len(result) == 2 and not result[1]:
+            # Auto-approval creation failed (returned False)
+            auto_approval_failed_count += 1
+
     # Count results
     approved_count = sum(
         1
@@ -157,30 +301,53 @@ async def process_review_action(
         if review.status == ReviewStatus.REJECTED
     )
 
-    # Resume execution if we processed some reviews
+    # Resume execution only if ALL pending reviews for this execution have been processed
     if updated_reviews:
-        # Get graph execution ID from any processed review
-        first_review = next(iter(updated_reviews.values()))
-        graph_exec_id = first_review.graph_exec_id
-
-        # Check if any pending reviews remain for this execution
         still_has_pending = await has_pending_reviews_for_graph_exec(graph_exec_id)
 
         if not still_has_pending:
-            # Resume execution
+            # Get the graph_id from any processed review
+            first_review = next(iter(updated_reviews.values()))
+
             try:
+                # Fetch user and settings to build complete execution context
+                user = await get_user_by_id(user_id)
+                settings = await get_graph_settings(
+                    user_id=user_id, graph_id=first_review.graph_id
+                )
+
+                # Preserve user's timezone preference when resuming execution
+                user_timezone = (
+                    user.timezone if user.timezone != USER_TIMEZONE_NOT_SET else "UTC"
+                )
+
+                execution_context = ExecutionContext(
+                    human_in_the_loop_safe_mode=settings.human_in_the_loop_safe_mode,
+                    sensitive_action_safe_mode=settings.sensitive_action_safe_mode,
+                    user_timezone=user_timezone,
+                )
+
                 await add_graph_execution(
                     graph_id=first_review.graph_id,
                     user_id=user_id,
                     graph_exec_id=graph_exec_id,
+                    execution_context=execution_context,
                 )
                 logger.info(f"Resumed execution {graph_exec_id}")
             except Exception as e:
                 logger.error(f"Failed to resume execution {graph_exec_id}: {str(e)}")
 
+    # Build error message if auto-approvals failed
+    error_message = None
+    if auto_approval_failed_count > 0:
+        error_message = (
+            f"{auto_approval_failed_count} auto-approval setting(s) could not be saved. "
+            f"You may need to manually approve these reviews in future executions."
+        )
+
     return ReviewResponse(
         approved_count=approved_count,
         rejected_count=rejected_count,
-        failed_count=0,
-        error=None,
+        failed_count=auto_approval_failed_count,
+        error=error_message,
     )

autogpt_platform/backend/backend/api/features/library/db.py

@@ -19,9 +19,12 @@ from backend.data.graph import GraphSettings
 from backend.data.includes import AGENT_PRESET_INCLUDE, library_agent_include
 from backend.data.model import CredentialsMetaInput
 from backend.integrations.creds_manager import IntegrationCredentialsManager
-from backend.integrations.webhooks.graph_lifecycle_hooks import on_graph_activate
+from backend.integrations.webhooks.graph_lifecycle_hooks import (
+    on_graph_activate,
+    on_graph_deactivate,
+)
 from backend.util.clients import get_scheduler_client
-from backend.util.exceptions import DatabaseError, NotFoundError
+from backend.util.exceptions import DatabaseError, InvalidInputError, NotFoundError
 from backend.util.json import SafeJson
 from backend.util.models import Pagination
 from backend.util.settings import Config
@@ -39,6 +42,7 @@ async def list_library_agents(
     sort_by: library_model.LibraryAgentSort = library_model.LibraryAgentSort.UPDATED_AT,
     page: int = 1,
     page_size: int = 50,
+    include_executions: bool = False,
 ) -> library_model.LibraryAgentResponse:
     """
     Retrieves a paginated list of LibraryAgent records for a given user.
@@ -49,6 +53,9 @@ async def list_library_agents(
         sort_by: Sorting field (createdAt, updatedAt, isFavorite, isCreatedByUser).
         page: Current page (1-indexed).
         page_size: Number of items per page.
+        include_executions: Whether to include execution data for status calculation.
+            Defaults to False for performance (UI fetches status separately).
+            Set to True when accurate status/metrics are needed (e.g., agent generator).
 
     Returns:
         A LibraryAgentResponse containing the list of agents and pagination details.
@@ -64,11 +71,11 @@ async def list_library_agents(
 
     if page < 1 or page_size < 1:
         logger.warning(f"Invalid pagination: page={page}, page_size={page_size}")
-        raise DatabaseError("Invalid pagination input")
+        raise InvalidInputError("Invalid pagination input")
 
     if search_term and len(search_term.strip()) > 100:
         logger.warning(f"Search term too long: {repr(search_term)}")
-        raise DatabaseError("Search term is too long")
+        raise InvalidInputError("Search term is too long")
 
     where_clause: prisma.types.LibraryAgentWhereInput = {
         "userId": user_id,
@@ -76,7 +83,6 @@ async def list_library_agents(
         "isArchived": False,
     }
 
-    # Build search filter if applicable
     if search_term:
         where_clause["OR"] = [
             {
@@ -93,7 +99,6 @@ async def list_library_agents(
             },
         ]
 
-    # Determine sorting
     order_by: prisma.types.LibraryAgentOrderByInput | None = None
 
     if sort_by == library_model.LibraryAgentSort.CREATED_AT:
@@ -105,7 +110,7 @@ async def list_library_agents(
         library_agents = await prisma.models.LibraryAgent.prisma().find_many(
             where=where_clause,
             include=library_agent_include(
-                user_id, include_nodes=False, include_executions=False
+                user_id, include_nodes=False, include_executions=include_executions
             ),
             order=order_by,
             skip=(page - 1) * page_size,
@@ -175,7 +180,7 @@ async def list_favorite_library_agents(
 
     if page < 1 or page_size < 1:
         logger.warning(f"Invalid pagination: page={page}, page_size={page_size}")
-        raise DatabaseError("Invalid pagination input")
+        raise InvalidInputError("Invalid pagination input")
 
     where_clause: prisma.types.LibraryAgentWhereInput = {
         "userId": user_id,
@@ -369,7 +374,7 @@ async def get_library_agent_by_graph_id(
 
 
 async def add_generated_agent_image(
-    graph: graph_db.BaseGraph,
+    graph: graph_db.GraphBaseMeta,
     user_id: str,
     library_agent_id: str,
 ) -> Optional[prisma.models.LibraryAgent]:
@@ -535,6 +540,92 @@ async def update_agent_version_in_library(
     return library_model.LibraryAgent.from_db(lib)
 
 
+async def create_graph_in_library(
+    graph: graph_db.Graph,
+    user_id: str,
+) -> tuple[graph_db.GraphModel, library_model.LibraryAgent]:
+    """Create a new graph and add it to the user's library."""
+    graph.version = 1
+    graph_model = graph_db.make_graph_model(graph, user_id)
+    graph_model.reassign_ids(user_id=user_id, reassign_graph_id=True)
+
+    created_graph = await graph_db.create_graph(graph_model, user_id)
+
+    library_agents = await create_library_agent(
+        graph=created_graph,
+        user_id=user_id,
+        sensitive_action_safe_mode=True,
+        create_library_agents_for_sub_graphs=False,
+    )
+
+    if created_graph.is_active:
+        created_graph = await on_graph_activate(created_graph, user_id=user_id)
+
+    return created_graph, library_agents[0]
+
+
+async def update_graph_in_library(
+    graph: graph_db.Graph,
+    user_id: str,
+) -> tuple[graph_db.GraphModel, library_model.LibraryAgent]:
+    """Create a new version of an existing graph and update the library entry."""
+    existing_versions = await graph_db.get_graph_all_versions(graph.id, user_id)
+    current_active_version = (
+        next((v for v in existing_versions if v.is_active), None)
+        if existing_versions
+        else None
+    )
+    graph.version = (
+        max(v.version for v in existing_versions) + 1 if existing_versions else 1
+    )
+
+    graph_model = graph_db.make_graph_model(graph, user_id)
+    graph_model.reassign_ids(user_id=user_id, reassign_graph_id=False)
+
+    created_graph = await graph_db.create_graph(graph_model, user_id)
+
+    library_agent = await get_library_agent_by_graph_id(user_id, created_graph.id)
+    if not library_agent:
+        raise NotFoundError(f"Library agent not found for graph {created_graph.id}")
+
+    library_agent = await update_library_agent_version_and_settings(
+        user_id, created_graph
+    )
+
+    if created_graph.is_active:
+        created_graph = await on_graph_activate(created_graph, user_id=user_id)
+        await graph_db.set_graph_active_version(
+            graph_id=created_graph.id,
+            version=created_graph.version,
+            user_id=user_id,
+        )
+        if current_active_version:
+            await on_graph_deactivate(current_active_version, user_id=user_id)
+
+    return created_graph, library_agent
+
+
+async def update_library_agent_version_and_settings(
+    user_id: str, agent_graph: graph_db.GraphModel
+) -> library_model.LibraryAgent:
+    """Update library agent to point to new graph version and sync settings."""
+    library = await update_agent_version_in_library(
+        user_id, agent_graph.id, agent_graph.version
+    )
+    updated_settings = GraphSettings.from_graph(
+        graph=agent_graph,
+        hitl_safe_mode=library.settings.human_in_the_loop_safe_mode,
+        sensitive_action_safe_mode=library.settings.sensitive_action_safe_mode,
+    )
+    if updated_settings != library.settings:
+        library = await update_library_agent(
+            library_agent_id=library.id,
+            user_id=user_id,
+            settings=updated_settings,
+        )
+    return library
+
+
 async def update_library_agent(
     library_agent_id: str,
     user_id: str,
@@ -583,7 +674,13 @@ async def update_library_agent(
             )
         update_fields["isDeleted"] = is_deleted
     if settings is not None:
-        update_fields["settings"] = SafeJson(settings.model_dump())
+        existing_agent = await get_library_agent(id=library_agent_id, user_id=user_id)
+        current_settings_dict = (
+            existing_agent.settings.model_dump() if existing_agent.settings else {}
+        )
+        new_settings = settings.model_dump(exclude_unset=True)
+        merged_settings = {**current_settings_dict, **new_settings}
+        update_fields["settings"] = SafeJson(merged_settings)
 
     try:
         # If graph_version is provided, update to that specific version

autogpt_platform/backend/backend/api/features/library/model.py

@@ -9,6 +9,7 @@ import pydantic
 from backend.data.block import BlockInput
 from backend.data.graph import GraphModel, GraphSettings, GraphTriggerInfo
 from backend.data.model import CredentialsMetaInput, is_credentials_field_name
+from backend.util.json import loads as json_loads
 from backend.util.models import Pagination
 
 if TYPE_CHECKING:
@@ -16,10 +17,10 @@ if TYPE_CHECKING:
 
 
 class LibraryAgentStatus(str, Enum):
-    COMPLETED = "COMPLETED"  # All runs completed
-    HEALTHY = "HEALTHY"  # Agent is running (not all runs have completed)
-    WAITING = "WAITING"  # Agent is queued or waiting to start
-    ERROR = "ERROR"  # Agent is in an error state
+    COMPLETED = "COMPLETED"
+    HEALTHY = "HEALTHY"
+    WAITING = "WAITING"
+    ERROR = "ERROR"
 
 
 class MarketplaceListingCreator(pydantic.BaseModel):
@@ -39,6 +40,30 @@ class MarketplaceListing(pydantic.BaseModel):
     creator: MarketplaceListingCreator
 
 
+class RecentExecution(pydantic.BaseModel):
+    """Summary of a recent execution for quality assessment.
+
+    Used by the LLM to understand the agent's recent performance with specific examples
+    rather than just aggregate statistics.
+    """
+
+    status: str
+    correctness_score: float | None = None
+    activity_summary: str | None = None
+
+
+def _parse_settings(settings: dict | str | None) -> GraphSettings:
+    """Parse settings from database, handling both dict and string formats."""
+    if settings is None:
+        return GraphSettings()
+    try:
+        if isinstance(settings, str):
+            settings = json_loads(settings)
+        return GraphSettings.model_validate(settings)
+    except Exception:
+        return GraphSettings()
+
+
 class LibraryAgent(pydantic.BaseModel):
     """
     Represents an agent in the library, including metadata for display and
@@ -48,7 +73,7 @@ class LibraryAgent(pydantic.BaseModel):
     id: str
     graph_id: str
     graph_version: int
-    owner_user_id: str  # ID of user who owns/created this agent graph
+    owner_user_id: str
 
     image_url: str | None
 
@@ -64,7 +89,7 @@ class LibraryAgent(pydantic.BaseModel):
     description: str
     instructions: str | None = None
 
-    input_schema: dict[str, Any]  # Should be BlockIOObjectSubSchema in frontend
+    input_schema: dict[str, Any]
     output_schema: dict[str, Any]
     credentials_input_schema: dict[str, Any] | None = pydantic.Field(
         description="Input schema for credentials required by the agent",
@@ -81,25 +106,19 @@ class LibraryAgent(pydantic.BaseModel):
     )
     trigger_setup_info: Optional[GraphTriggerInfo] = None
 
-    # Indicates whether there's a new output (based on recent runs)
     new_output: bool
-
-    # Whether the user can access the underlying graph
+    execution_count: int = 0
+    success_rate: float | None = None
+    avg_correctness_score: float | None = None
+    recent_executions: list[RecentExecution] = pydantic.Field(
+        default_factory=list,
+        description="List of recent executions with status, score, and summary",
+    )
     can_access_graph: bool
-
-    # Indicates if this agent is the latest version
     is_latest_version: bool
-
-    # Whether the agent is marked as favorite by the user
     is_favorite: bool
-
-    # Recommended schedule cron (from marketplace agents)
     recommended_schedule_cron: str | None = None
-
-    # User-specific settings for this library agent
     settings: GraphSettings = pydantic.Field(default_factory=GraphSettings)
-
-    # Marketplace listing information if the agent has been published
     marketplace_listing: Optional["MarketplaceListing"] = None
 
     @staticmethod
@@ -123,7 +142,6 @@ class LibraryAgent(pydantic.BaseModel):
         agent_updated_at = agent.AgentGraph.updatedAt
         lib_agent_updated_at = agent.updatedAt
 
-        # Compute updated_at as the latest between library agent and graph
         updated_at = (
             max(agent_updated_at, lib_agent_updated_at)
             if agent_updated_at
@@ -136,7 +154,6 @@ class LibraryAgent(pydantic.BaseModel):
             creator_name = agent.Creator.name or "Unknown"
             creator_image_url = agent.Creator.avatarUrl or ""
 
-        # Logic to calculate status and new_output
         week_ago = datetime.datetime.now(datetime.timezone.utc) - datetime.timedelta(
             days=7
         )
@@ -145,13 +162,55 @@ class LibraryAgent(pydantic.BaseModel):
         status = status_result.status
         new_output = status_result.new_output
 
-        # Check if user can access the graph
-        can_access_graph = agent.AgentGraph.userId == agent.userId
+        execution_count = len(executions)
+        success_rate: float | None = None
+        avg_correctness_score: float | None = None
+        if execution_count > 0:
+            success_count = sum(
+                1
+                for e in executions
+                if e.executionStatus == prisma.enums.AgentExecutionStatus.COMPLETED
+            )
+            success_rate = (success_count / execution_count) * 100
 
-        # Hard-coded to True until a method to check is implemented
+            correctness_scores = []
+            for e in executions:
+                if e.stats and isinstance(e.stats, dict):
+                    score = e.stats.get("correctness_score")
+                    if score is not None and isinstance(score, (int, float)):
+                        correctness_scores.append(float(score))
+            if correctness_scores:
+                avg_correctness_score = sum(correctness_scores) / len(
+                    correctness_scores
+                )
+
+        recent_executions: list[RecentExecution] = []
+        for e in executions:
+            exec_score: float | None = None
+            exec_summary: str | None = None
+            if e.stats and isinstance(e.stats, dict):
+                score = e.stats.get("correctness_score")
+                if score is not None and isinstance(score, (int, float)):
+                    exec_score = float(score)
+                summary = e.stats.get("activity_status")
+                if summary is not None and isinstance(summary, str):
+                    exec_summary = summary
+            exec_status = (
+                e.executionStatus.value
+                if hasattr(e.executionStatus, "value")
+                else str(e.executionStatus)
+            )
+            recent_executions.append(
+                RecentExecution(
+                    status=exec_status,
+                    correctness_score=exec_score,
+                    activity_summary=exec_summary,
+                )
+            )
+
+        can_access_graph = agent.AgentGraph.userId == agent.userId
         is_latest_version = True
 
-        # Build marketplace_listing if available
         marketplace_listing_data = None
         if store_listing and store_listing.ActiveVersion and profile:
             creator_data = MarketplaceListingCreator(
@@ -190,11 +249,15 @@ class LibraryAgent(pydantic.BaseModel):
             has_sensitive_action=graph.has_sensitive_action,
             trigger_setup_info=graph.trigger_setup_info,
             new_output=new_output,
+            execution_count=execution_count,
+            success_rate=success_rate,
+            avg_correctness_score=avg_correctness_score,
+            recent_executions=recent_executions,
             can_access_graph=can_access_graph,
             is_latest_version=is_latest_version,
             is_favorite=agent.isFavorite,
             recommended_schedule_cron=agent.AgentGraph.recommendedScheduleCron,
-            settings=GraphSettings.model_validate(agent.settings),
+            settings=_parse_settings(agent.settings),
             marketplace_listing=marketplace_listing_data,
         )
 
@@ -220,18 +283,15 @@ def _calculate_agent_status(
     if not executions:
         return AgentStatusResult(status=LibraryAgentStatus.COMPLETED, new_output=False)
 
-    # Track how many times each execution status appears
     status_counts = {status: 0 for status in prisma.enums.AgentExecutionStatus}
     new_output = False
 
     for execution in executions:
-        # Check if there's a completed run more recent than `recent_threshold`
         if execution.createdAt >= recent_threshold:
             if execution.executionStatus == prisma.enums.AgentExecutionStatus.COMPLETED:
                 new_output = True
         status_counts[execution.executionStatus] += 1
 
-    # Determine the final status based on counts
     if status_counts[prisma.enums.AgentExecutionStatus.FAILED] > 0:
         return AgentStatusResult(status=LibraryAgentStatus.ERROR, new_output=new_output)
     elif status_counts[prisma.enums.AgentExecutionStatus.QUEUED] > 0:

autogpt_platform/backend/backend/api/features/library/routes/agents.py

@@ -1,4 +1,3 @@
-import logging
 from typing import Literal, Optional
 
 import autogpt_libs.auth as autogpt_auth_lib
@@ -6,15 +5,11 @@ from fastapi import APIRouter, Body, HTTPException, Query, Security, status
 from fastapi.responses import Response
 from prisma.enums import OnboardingStep
 
-import backend.api.features.store.exceptions as store_exceptions
 from backend.data.onboarding import complete_onboarding_step
-from backend.util.exceptions import DatabaseError, NotFoundError
 
 from .. import db as library_db
 from .. import model as library_model
 
-logger = logging.getLogger(__name__)
-
 router = APIRouter(
     prefix="/agents",
     tags=["library", "private"],
@@ -26,10 +21,6 @@ router = APIRouter(
     "",
     summary="List Library Agents",
     response_model=library_model.LibraryAgentResponse,
-    responses={
-        200: {"description": "List of library agents"},
-        500: {"description": "Server error", "content": {"application/json": {}}},
-    },
 )
 async def list_library_agents(
     user_id: str = Security(autogpt_auth_lib.get_user_id),
@@ -53,43 +44,19 @@ async def list_library_agents(
 ) -> library_model.LibraryAgentResponse:
     """
     Get all agents in the user's library (both created and saved).
-
-    Args:
-        user_id: ID of the authenticated user.
-        search_term: Optional search term to filter agents by name/description.
-        filter_by: List of filters to apply (favorites, created by user).
-        sort_by: List of sorting criteria (created date, updated date).
-        page: Page number to retrieve.
-        page_size: Number of agents per page.
-
-    Returns:
-        A LibraryAgentResponse containing agents and pagination metadata.
-
-    Raises:
-        HTTPException: If a server/database error occurs.
     """
-    try:
-        return await library_db.list_library_agents(
-            user_id=user_id,
-            search_term=search_term,
-            sort_by=sort_by,
-            page=page,
-            page_size=page_size,
-        )
-    except Exception as e:
-        logger.error(f"Could not list library agents for user #{user_id}: {e}")
-        raise HTTPException(
-            status_code=status.HTTP_500_INTERNAL_SERVER_ERROR,
-            detail=str(e),
-        ) from e
+    return await library_db.list_library_agents(
+        user_id=user_id,
+        search_term=search_term,
+        sort_by=sort_by,
+        page=page,
+        page_size=page_size,
+    )
 
 
 @router.get(
     "/favorites",
     summary="List Favorite Library Agents",
-    responses={
-        500: {"description": "Server error", "content": {"application/json": {}}},
-    },
 )
 async def list_favorite_library_agents(
     user_id: str = Security(autogpt_auth_lib.get_user_id),
@@ -106,30 +73,12 @@ async def list_favorite_library_agents(
 ) -> library_model.LibraryAgentResponse:
     """
     Get all favorite agents in the user's library.
-
-    Args:
-        user_id: ID of the authenticated user.
-        page: Page number to retrieve.
-        page_size: Number of agents per page.
-
-    Returns:
-        A LibraryAgentResponse containing favorite agents and pagination metadata.
-
-    Raises:
-        HTTPException: If a server/database error occurs.
     """
-    try:
-        return await library_db.list_favorite_library_agents(
-            user_id=user_id,
-            page=page,
-            page_size=page_size,
-        )
-    except Exception as e:
-        logger.error(f"Could not list favorite library agents for user #{user_id}: {e}")
-        raise HTTPException(
-            status_code=status.HTTP_500_INTERNAL_SERVER_ERROR,
-            detail=str(e),
-        ) from e
+    return await library_db.list_favorite_library_agents(
+        user_id=user_id,
+        page=page,
+        page_size=page_size,
+    )
 
 
 @router.get("/{library_agent_id}", summary="Get Library Agent")
@@ -162,10 +111,6 @@ async def get_library_agent_by_graph_id(
     summary="Get Agent By Store ID",
     tags=["store", "library"],
     response_model=library_model.LibraryAgent | None,
-    responses={
-        200: {"description": "Library agent found"},
-        404: {"description": "Agent not found"},
-    },
 )
 async def get_library_agent_by_store_listing_version_id(
     store_listing_version_id: str,
@@ -174,32 +119,15 @@ async def get_library_agent_by_store_listing_version_id(
     """
     Get Library Agent from Store Listing Version ID.
     """
-    try:
-        return await library_db.get_library_agent_by_store_version_id(
-            store_listing_version_id, user_id
-        )
-    except NotFoundError as e:
-        raise HTTPException(
-            status_code=status.HTTP_404_NOT_FOUND,
-            detail=str(e),
-        )
-    except Exception as e:
-        logger.error(f"Could not fetch library agent from store version ID: {e}")
-        raise HTTPException(
-            status_code=status.HTTP_500_INTERNAL_SERVER_ERROR,
-            detail=str(e),
-        ) from e
+    return await library_db.get_library_agent_by_store_version_id(
+        store_listing_version_id, user_id
+    )
 
 
 @router.post(
     "",
     summary="Add Marketplace Agent",
     status_code=status.HTTP_201_CREATED,
-    responses={
-        201: {"description": "Agent added successfully"},
-        404: {"description": "Store listing version not found"},
-        500: {"description": "Server error"},
-    },
 )
 async def add_marketplace_agent_to_library(
     store_listing_version_id: str = Body(embed=True),
@@ -210,59 +138,19 @@ async def add_marketplace_agent_to_library(
 ) -> library_model.LibraryAgent:
     """
     Add an agent from the marketplace to the user's library.
-
-    Args:
-        store_listing_version_id: ID of the store listing version to add.
-        user_id: ID of the authenticated user.
-
-    Returns:
-        library_model.LibraryAgent: Agent added to the library
-
-    Raises:
-        HTTPException(404): If the listing version is not found.
-        HTTPException(500): If a server/database error occurs.
     """
-    try:
-        agent = await library_db.add_store_agent_to_library(
-            store_listing_version_id=store_listing_version_id,
-            user_id=user_id,
-        )
-        if source != "onboarding":
-            await complete_onboarding_step(
-                user_id, OnboardingStep.MARKETPLACE_ADD_AGENT
-            )
-        return agent
-
-    except store_exceptions.AgentNotFoundError as e:
-        logger.warning(
-            f"Could not find store listing version {store_listing_version_id} "
-            "to add to library"
-        )
-        raise HTTPException(status_code=status.HTTP_404_NOT_FOUND, detail=str(e))
-    except DatabaseError as e:
-        logger.error(f"Database error while adding agent to library: {e}", e)
-        raise HTTPException(
-            status_code=status.HTTP_500_INTERNAL_SERVER_ERROR,
-            detail={"message": str(e), "hint": "Inspect DB logs for details."},
-        ) from e
-    except Exception as e:
-        logger.error(f"Unexpected error while adding agent to library: {e}")
-        raise HTTPException(
-            status_code=status.HTTP_500_INTERNAL_SERVER_ERROR,
-            detail={
-                "message": str(e),
-                "hint": "Check server logs for more information.",
-            },
-        ) from e
+    agent = await library_db.add_store_agent_to_library(
+        store_listing_version_id=store_listing_version_id,
+        user_id=user_id,
+    )
+    if source != "onboarding":
+        await complete_onboarding_step(user_id, OnboardingStep.MARKETPLACE_ADD_AGENT)
+    return agent
 
 
 @router.patch(
     "/{library_agent_id}",
     summary="Update Library Agent",
-    responses={
-        200: {"description": "Agent updated successfully"},
-        500: {"description": "Server error"},
-    },
 )
 async def update_library_agent(
     library_agent_id: str,
@@ -271,52 +159,21 @@ async def update_library_agent(
 ) -> library_model.LibraryAgent:
     """
     Update the library agent with the given fields.
-
-    Args:
-        library_agent_id: ID of the library agent to update.
-        payload: Fields to update (auto_update_version, is_favorite, etc.).
-        user_id: ID of the authenticated user.
-
-    Raises:
-        HTTPException(500): If a server/database error occurs.
     """
-    try:
-        return await library_db.update_library_agent(
-            library_agent_id=library_agent_id,
-            user_id=user_id,
-            auto_update_version=payload.auto_update_version,
-            graph_version=payload.graph_version,
-            is_favorite=payload.is_favorite,
-            is_archived=payload.is_archived,
-            settings=payload.settings,
-        )
-    except NotFoundError as e:
-        raise HTTPException(
-            status_code=status.HTTP_404_NOT_FOUND,
-            detail=str(e),
-        ) from e
-    except DatabaseError as e:
-        logger.error(f"Database error while updating library agent: {e}")
-        raise HTTPException(
-            status_code=status.HTTP_500_INTERNAL_SERVER_ERROR,
-            detail={"message": str(e), "hint": "Verify DB connection."},
-        ) from e
-    except Exception as e:
-        logger.error(f"Unexpected error while updating library agent: {e}")
-        raise HTTPException(
-            status_code=status.HTTP_500_INTERNAL_SERVER_ERROR,
-            detail={"message": str(e), "hint": "Check server logs."},
-        ) from e
+    return await library_db.update_library_agent(
+        library_agent_id=library_agent_id,
+        user_id=user_id,
+        auto_update_version=payload.auto_update_version,
+        graph_version=payload.graph_version,
+        is_favorite=payload.is_favorite,
+        is_archived=payload.is_archived,
+        settings=payload.settings,
+    )
 
 
 @router.delete(
     "/{library_agent_id}",
     summary="Delete Library Agent",
-    responses={
-        204: {"description": "Agent deleted successfully"},
-        404: {"description": "Agent not found"},
-        500: {"description": "Server error"},
-    },
 )
 async def delete_library_agent(
     library_agent_id: str,
@@ -324,28 +181,11 @@ async def delete_library_agent(
 ) -> Response:
     """
     Soft-delete the specified library agent.
-
-    Args:
-        library_agent_id: ID of the library agent to delete.
-        user_id: ID of the authenticated user.
-
-    Returns:
-        204 No Content if successful.
-
-    Raises:
-        HTTPException(404): If the agent does not exist.
-        HTTPException(500): If a server/database error occurs.
     """
-    try:
-        await library_db.delete_library_agent(
-            library_agent_id=library_agent_id, user_id=user_id
-        )
-        return Response(status_code=status.HTTP_204_NO_CONTENT)
-    except NotFoundError as e:
-        raise HTTPException(
-            status_code=status.HTTP_404_NOT_FOUND,
-            detail=str(e),
-        ) from e
+    await library_db.delete_library_agent(
+        library_agent_id=library_agent_id, user_id=user_id
+    )
+    return Response(status_code=status.HTTP_204_NO_CONTENT)
 
 
 @router.post("/{library_agent_id}/fork", summary="Fork Library Agent")

autogpt_platform/backend/backend/api/features/library/routes_test.py

@@ -118,21 +118,6 @@ async def test_get_library_agents_success(
     )
 
 
-def test_get_library_agents_error(mocker: pytest_mock.MockFixture, test_user_id: str):
-    mock_db_call = mocker.patch("backend.api.features.library.db.list_library_agents")
-    mock_db_call.side_effect = Exception("Test error")
-
-    response = client.get("/agents?search_term=test")
-    assert response.status_code == 500
-    mock_db_call.assert_called_once_with(
-        user_id=test_user_id,
-        search_term="test",
-        sort_by=library_model.LibraryAgentSort.UPDATED_AT,
-        page=1,
-        page_size=15,
-    )
-
-
 @pytest.mark.asyncio
 async def test_get_favorite_library_agents_success(
     mocker: pytest_mock.MockFixture,
@@ -190,23 +175,6 @@ async def test_get_favorite_library_agents_success(
     )
 
 
-def test_get_favorite_library_agents_error(
-    mocker: pytest_mock.MockFixture, test_user_id: str
-):
-    mock_db_call = mocker.patch(
-        "backend.api.features.library.db.list_favorite_library_agents"
-    )
-    mock_db_call.side_effect = Exception("Test error")
-
-    response = client.get("/agents/favorites")
-    assert response.status_code == 500
-    mock_db_call.assert_called_once_with(
-        user_id=test_user_id,
-        page=1,
-        page_size=15,
-    )
-
-
 def test_add_agent_to_library_success(
     mocker: pytest_mock.MockFixture, test_user_id: str
 ):
@@ -258,19 +226,3 @@ def test_add_agent_to_library_success(
         store_listing_version_id="test-version-id", user_id=test_user_id
     )
     mock_complete_onboarding.assert_awaited_once()
-
-
-def test_add_agent_to_library_error(mocker: pytest_mock.MockFixture, test_user_id: str):
-    mock_db_call = mocker.patch(
-        "backend.api.features.library.db.add_store_agent_to_library"
-    )
-    mock_db_call.side_effect = Exception("Test error")
-
-    response = client.post(
-        "/agents", json={"store_listing_version_id": "test-version-id"}
-    )
-    assert response.status_code == 500
-    assert "detail" in response.json()  # Verify error response structure
-    mock_db_call.assert_called_once_with(
-        store_listing_version_id="test-version-id", user_id=test_user_id
-    )

autogpt_platform/backend/backend/api/features/oauth_test.py

@@ -20,6 +20,7 @@ from typing import AsyncGenerator
 
 import httpx
 import pytest
+import pytest_asyncio
 from autogpt_libs.api_key.keysmith import APIKeySmith
 from prisma.enums import APIKeyPermission
 from prisma.models import OAuthAccessToken as PrismaOAuthAccessToken
@@ -38,13 +39,13 @@ keysmith = APIKeySmith()
 # ============================================================================
 
 
-@pytest.fixture
+@pytest.fixture(scope="session")
 def test_user_id() -> str:
     """Test user ID for OAuth tests."""
     return str(uuid.uuid4())
 
 
-@pytest.fixture
+@pytest_asyncio.fixture(scope="session", loop_scope="session")
 async def test_user(server, test_user_id: str):
     """Create a test user in the database."""
     await PrismaUser.prisma().create(
@@ -67,7 +68,7 @@ async def test_user(server, test_user_id: str):
     await PrismaUser.prisma().delete(where={"id": test_user_id})
 
 
-@pytest.fixture
+@pytest_asyncio.fixture
 async def test_oauth_app(test_user: str):
     """Create a test OAuth application in the database."""
     app_id = str(uuid.uuid4())
@@ -122,7 +123,7 @@ def pkce_credentials() -> tuple[str, str]:
     return generate_pkce()
 
 
-@pytest.fixture
+@pytest_asyncio.fixture
 async def client(server, test_user: str) -> AsyncGenerator[httpx.AsyncClient, None]:
     """
     Create an async HTTP client that talks directly to the FastAPI app.
@@ -287,7 +288,7 @@ async def test_authorize_invalid_client_returns_error(
     assert query_params["error"][0] == "invalid_client"
 
 
-@pytest.fixture
+@pytest_asyncio.fixture
 async def inactive_oauth_app(test_user: str):
     """Create an inactive test OAuth application in the database."""
     app_id = str(uuid.uuid4())
@@ -1004,7 +1005,7 @@ async def test_token_refresh_revoked(
     assert "revoked" in response.json()["detail"].lower()
 
 
-@pytest.fixture
+@pytest_asyncio.fixture
 async def other_oauth_app(test_user: str):
     """Create a second OAuth application for cross-app tests."""
     app_id = str(uuid.uuid4())

autogpt_platform/backend/backend/api/features/store/content_handlers.py

@@ -188,6 +188,10 @@ class BlockHandler(ContentHandler):
             try:
                 block_instance = block_cls()
 
+                # Skip disabled blocks - they shouldn't be indexed
+                if block_instance.disabled:
+                    continue
+
                 # Build searchable text from block metadata
                 parts = []
                 if hasattr(block_instance, "name") and block_instance.name:
@@ -248,12 +252,19 @@ class BlockHandler(ContentHandler):
         from backend.data.block import get_blocks
 
         all_blocks = get_blocks()
-        total_blocks = len(all_blocks)
+
+        # Filter out disabled blocks - they're not indexed
+        enabled_block_ids = [
+            block_id
+            for block_id, block_cls in all_blocks.items()
+            if not block_cls().disabled
+        ]
+        total_blocks = len(enabled_block_ids)
 
         if total_blocks == 0:
             return {"total": 0, "with_embeddings": 0, "without_embeddings": 0}
 
-        block_ids = list(all_blocks.keys())
+        block_ids = enabled_block_ids
         placeholders = ",".join([f"${i+1}" for i in range(len(block_ids))])
 
         embedded_result = await query_raw_with_schema(

autogpt_platform/backend/backend/api/features/store/content_handlers_test.py

@@ -81,6 +81,7 @@ async def test_block_handler_get_missing_items(mocker):
     mock_block_instance.name = "Calculator Block"
     mock_block_instance.description = "Performs calculations"
     mock_block_instance.categories = [MagicMock(value="MATH")]
+    mock_block_instance.disabled = False
     mock_block_instance.input_schema.model_json_schema.return_value = {
         "properties": {"expression": {"description": "Math expression to evaluate"}}
     }
@@ -116,11 +117,18 @@ async def test_block_handler_get_stats(mocker):
     """Test BlockHandler returns correct stats."""
     handler = BlockHandler()
 
-    # Mock get_blocks
+    # Mock get_blocks - each block class returns an instance with disabled=False
+    def make_mock_block_class():
+        mock_class = MagicMock()
+        mock_instance = MagicMock()
+        mock_instance.disabled = False
+        mock_class.return_value = mock_instance
+        return mock_class
+
     mock_blocks = {
-        "block-1": MagicMock(),
-        "block-2": MagicMock(),
-        "block-3": MagicMock(),
+        "block-1": make_mock_block_class(),
+        "block-2": make_mock_block_class(),
+        "block-3": make_mock_block_class(),
     }
 
     # Mock embedded count query (2 blocks have embeddings)
@@ -309,6 +317,7 @@ async def test_block_handler_handles_missing_attributes():
     mock_block_class = MagicMock()
     mock_block_instance = MagicMock()
     mock_block_instance.name = "Minimal Block"
+    mock_block_instance.disabled = False
     # No description, categories, or schema
     del mock_block_instance.description
     del mock_block_instance.categories
@@ -342,6 +351,7 @@ async def test_block_handler_skips_failed_blocks():
     good_instance.name = "Good Block"
     good_instance.description = "Works fine"
     good_instance.categories = []
+    good_instance.disabled = False
     good_block.return_value = good_instance
 
     bad_block = MagicMock()

autogpt_platform/backend/backend/api/features/store/db.py

@@ -1,7 +1,7 @@
 import asyncio
 import logging
 from datetime import datetime, timezone
-from typing import Any, Literal
+from typing import Any, Literal, overload
 
 import fastapi
 import prisma.enums
@@ -11,8 +11,8 @@ import prisma.types
 
 from backend.data.db import transaction
 from backend.data.graph import (
-    GraphMeta,
     GraphModel,
+    GraphModelWithoutNodes,
     get_graph,
     get_graph_as_admin,
     get_sub_graphs,
@@ -112,6 +112,7 @@ async def get_store_agents(
                             description=agent["description"],
                             runs=agent["runs"],
                             rating=agent["rating"],
+                            agent_graph_id=agent.get("agentGraphId", ""),
                         )
                         store_agents.append(store_agent)
                     except Exception as e:
@@ -170,6 +171,7 @@ async def get_store_agents(
                         description=agent.description,
                         runs=agent.runs,
                         rating=agent.rating,
+                        agent_graph_id=agent.agentGraphId,
                     )
                     # Add to the list only if creation was successful
                     store_agents.append(store_agent)
@@ -332,7 +334,22 @@ async def get_store_agent_details(
         raise DatabaseError("Failed to fetch agent details") from e
 
 
-async def get_available_graph(store_listing_version_id: str) -> GraphMeta:
+@overload
+async def get_available_graph(
+    store_listing_version_id: str, hide_nodes: Literal[False]
+) -> GraphModel: ...
+
+
+@overload
+async def get_available_graph(
+    store_listing_version_id: str, hide_nodes: Literal[True] = True
+) -> GraphModelWithoutNodes: ...
+
+
+async def get_available_graph(
+    store_listing_version_id: str,
+    hide_nodes: bool = True,
+) -> GraphModelWithoutNodes | GraphModel:
     try:
         # Get avaialble, non-deleted store listing version
         store_listing_version = (
@@ -342,7 +359,7 @@ async def get_available_graph(store_listing_version_id: str) -> GraphMeta:
                     "isAvailable": True,
                     "isDeleted": False,
                 },
-                include={"AgentGraph": {"include": {"Nodes": True}}},
+                include={"AgentGraph": {"include": AGENT_GRAPH_INCLUDE}},
             )
         )
 
@@ -352,7 +369,9 @@ async def get_available_graph(store_listing_version_id: str) -> GraphMeta:
                 detail=f"Store listing version {store_listing_version_id} not found",
             )
 
-        return GraphModel.from_db(store_listing_version.AgentGraph).meta()
+        return (GraphModelWithoutNodes if hide_nodes else GraphModel).from_db(
+            store_listing_version.AgentGraph
+        )
 
     except Exception as e:
         logger.error(f"Error getting agent: {e}")
@@ -1552,7 +1571,7 @@ async def review_store_submission(
 
                 # Generate embedding for approved listing (blocking - admin operation)
                 # Inside transaction: if embedding fails, entire transaction rolls back
-                embedding_success = await ensure_embedding(
+                await ensure_embedding(
                     version_id=store_listing_version_id,
                     name=store_listing_version.name,
                     description=store_listing_version.description,
@@ -1560,12 +1579,6 @@ async def review_store_submission(
                     categories=store_listing_version.categories or [],
                     tx=tx,
                 )
-                if not embedding_success:
-                    raise ValueError(
-                        f"Failed to generate embedding for listing {store_listing_version_id}. "
-                        "This is likely due to OpenAI API being unavailable. "
-                        "Please try again later or contact support if the issue persists."
-                    )
 
                 await prisma.models.StoreListing.prisma(tx).update(
                     where={"id": store_listing_version.StoreListing.id},

autogpt_platform/backend/backend/api/features/store/embeddings.py

@@ -21,7 +21,6 @@ from backend.util.json import dumps
 
 logger = logging.getLogger(__name__)
 
-
 # OpenAI embedding model configuration
 EMBEDDING_MODEL = "text-embedding-3-small"
 # Embedding dimension for the model above
@@ -63,49 +62,42 @@ def build_searchable_text(
     return " ".join(parts)
 
 
-async def generate_embedding(text: str) -> list[float] | None:
+async def generate_embedding(text: str) -> list[float]:
     """
     Generate embedding for text using OpenAI API.
 
-    Returns None if embedding generation fails.
-    Fail-fast: no retries to maintain consistency with approval flow.
+    Raises exceptions on failure - caller should handle.
     """
-    try:
-        client = get_openai_client()
-        if not client:
-            logger.error("openai_internal_api_key not set, cannot generate embedding")
-            return None
+    client = get_openai_client()
+    if not client:
+        raise RuntimeError("openai_internal_api_key not set, cannot generate embedding")
 
-        # Truncate text to token limit using tiktoken
-        # Character-based truncation is insufficient because token ratios vary by content type
-        enc = encoding_for_model(EMBEDDING_MODEL)
-        tokens = enc.encode(text)
-        if len(tokens) > EMBEDDING_MAX_TOKENS:
-            tokens = tokens[:EMBEDDING_MAX_TOKENS]
-            truncated_text = enc.decode(tokens)
-            logger.info(
-                f"Truncated text from {len(enc.encode(text))} to {len(tokens)} tokens"
-            )
-        else:
-            truncated_text = text
-
-        start_time = time.time()
-        response = await client.embeddings.create(
-            model=EMBEDDING_MODEL,
-            input=truncated_text,
-        )
-        latency_ms = (time.time() - start_time) * 1000
-
-        embedding = response.data[0].embedding
+    # Truncate text to token limit using tiktoken
+    # Character-based truncation is insufficient because token ratios vary by content type
+    enc = encoding_for_model(EMBEDDING_MODEL)
+    tokens = enc.encode(text)
+    if len(tokens) > EMBEDDING_MAX_TOKENS:
+        tokens = tokens[:EMBEDDING_MAX_TOKENS]
+        truncated_text = enc.decode(tokens)
         logger.info(
-            f"Generated embedding: {len(embedding)} dims, "
-            f"{len(tokens)} tokens, {latency_ms:.0f}ms"
+            f"Truncated text from {len(enc.encode(text))} to {len(tokens)} tokens"
         )
-        return embedding
+    else:
+        truncated_text = text
 
-    except Exception as e:
-        logger.error(f"Failed to generate embedding: {e}")
-        return None
+    start_time = time.time()
+    response = await client.embeddings.create(
+        model=EMBEDDING_MODEL,
+        input=truncated_text,
+    )
+    latency_ms = (time.time() - start_time) * 1000
+
+    embedding = response.data[0].embedding
+    logger.info(
+        f"Generated embedding: {len(embedding)} dims, "
+        f"{len(tokens)} tokens, {latency_ms:.0f}ms"
+    )
+    return embedding
 
 
 async def store_embedding(
@@ -144,48 +136,45 @@ async def store_content_embedding(
 
     New function for unified content embedding storage.
     Uses raw SQL since Prisma doesn't natively support pgvector.
+
+    Raises exceptions on failure - caller should handle.
     """
-    try:
-        client = tx if tx else prisma.get_client()
+    client = tx if tx else prisma.get_client()
 
-        # Convert embedding to PostgreSQL vector format
-        embedding_str = embedding_to_vector_string(embedding)
-        metadata_json = dumps(metadata or {})
+    # Convert embedding to PostgreSQL vector format
+    embedding_str = embedding_to_vector_string(embedding)
+    metadata_json = dumps(metadata or {})
 
-        # Upsert the embedding
-        # WHERE clause in DO UPDATE prevents PostgreSQL 15 bug with NULLS NOT DISTINCT
-        # Use unqualified ::vector - pgvector is in search_path on all environments
-        await execute_raw_with_schema(
-            """
-            INSERT INTO {schema_prefix}"UnifiedContentEmbedding" (
-                "id", "contentType", "contentId", "userId", "embedding", "searchableText", "metadata", "createdAt", "updatedAt"
-            )
-            VALUES (gen_random_uuid()::text, $1::{schema_prefix}"ContentType", $2, $3, $4::vector, $5, $6::jsonb, NOW(), NOW())
-            ON CONFLICT ("contentType", "contentId", "userId")
-            DO UPDATE SET
-                "embedding" = $4::vector,
-                "searchableText" = $5,
-                "metadata" = $6::jsonb,
-                "updatedAt" = NOW()
-            WHERE {schema_prefix}"UnifiedContentEmbedding"."contentType" = $1::{schema_prefix}"ContentType"
-                AND {schema_prefix}"UnifiedContentEmbedding"."contentId" = $2
-                AND ({schema_prefix}"UnifiedContentEmbedding"."userId" = $3 OR ($3 IS NULL AND {schema_prefix}"UnifiedContentEmbedding"."userId" IS NULL))
-            """,
-            content_type,
-            content_id,
-            user_id,
-            embedding_str,
-            searchable_text,
-            metadata_json,
-            client=client,
+    # Upsert the embedding
+    # WHERE clause in DO UPDATE prevents PostgreSQL 15 bug with NULLS NOT DISTINCT
+    # Use unqualified ::vector - pgvector is in search_path on all environments
+    await execute_raw_with_schema(
+        """
+        INSERT INTO {schema_prefix}"UnifiedContentEmbedding" (
+            "id", "contentType", "contentId", "userId", "embedding", "searchableText", "metadata", "createdAt", "updatedAt"
         )
+        VALUES (gen_random_uuid()::text, $1::{schema_prefix}"ContentType", $2, $3, $4::vector, $5, $6::jsonb, NOW(), NOW())
+        ON CONFLICT ("contentType", "contentId", "userId")
+        DO UPDATE SET
+            "embedding" = $4::vector,
+            "searchableText" = $5,
+            "metadata" = $6::jsonb,
+            "updatedAt" = NOW()
+        WHERE {schema_prefix}"UnifiedContentEmbedding"."contentType" = $1::{schema_prefix}"ContentType"
+            AND {schema_prefix}"UnifiedContentEmbedding"."contentId" = $2
+            AND ({schema_prefix}"UnifiedContentEmbedding"."userId" = $3 OR ($3 IS NULL AND {schema_prefix}"UnifiedContentEmbedding"."userId" IS NULL))
+        """,
+        content_type,
+        content_id,
+        user_id,
+        embedding_str,
+        searchable_text,
+        metadata_json,
+        client=client,
+    )
 
-        logger.info(f"Stored embedding for {content_type}:{content_id}")
-        return True
-
-    except Exception as e:
-        logger.error(f"Failed to store embedding for {content_type}:{content_id}: {e}")
-        return False
+    logger.info(f"Stored embedding for {content_type}:{content_id}")
+    return True
 
 
 async def get_embedding(version_id: str) -> dict[str, Any] | None:
@@ -217,34 +206,31 @@ async def get_content_embedding(
 
     New function for unified content embedding retrieval.
     Returns dict with contentType, contentId, embedding, timestamps or None if not found.
+
+    Raises exceptions on failure - caller should handle.
     """
-    try:
-        result = await query_raw_with_schema(
-            """
-            SELECT
-                "contentType",
-                "contentId",
-                "userId",
-                "embedding"::text as "embedding",
-                "searchableText",
-                "metadata",
-                "createdAt",
-                "updatedAt"
-            FROM {schema_prefix}"UnifiedContentEmbedding"
-            WHERE "contentType" = $1::{schema_prefix}"ContentType" AND "contentId" = $2 AND ("userId" = $3 OR ($3 IS NULL AND "userId" IS NULL))
-            """,
-            content_type,
-            content_id,
-            user_id,
-        )
+    result = await query_raw_with_schema(
+        """
+        SELECT
+            "contentType",
+            "contentId",
+            "userId",
+            "embedding"::text as "embedding",
+            "searchableText",
+            "metadata",
+            "createdAt",
+            "updatedAt"
+        FROM {schema_prefix}"UnifiedContentEmbedding"
+        WHERE "contentType" = $1::{schema_prefix}"ContentType" AND "contentId" = $2 AND ("userId" = $3 OR ($3 IS NULL AND "userId" IS NULL))
+        """,
+        content_type,
+        content_id,
+        user_id,
+    )
 
-        if result and len(result) > 0:
-            return result[0]
-        return None
-
-    except Exception as e:
-        logger.error(f"Failed to get embedding for {content_type}:{content_id}: {e}")
-        return None
+    if result and len(result) > 0:
+        return result[0]
+    return None
 
 
 async def ensure_embedding(
@@ -272,46 +258,38 @@ async def ensure_embedding(
         tx: Optional transaction client
 
     Returns:
-        True if embedding exists/was created, False on failure
+        True if embedding exists/was created
+
+    Raises exceptions on failure - caller should handle.
     """
-    try:
-        # Check if embedding already exists
-        if not force:
-            existing = await get_embedding(version_id)
-            if existing and existing.get("embedding"):
-                logger.debug(f"Embedding for version {version_id} already exists")
-                return True
+    # Check if embedding already exists
+    if not force:
+        existing = await get_embedding(version_id)
+        if existing and existing.get("embedding"):
+            logger.debug(f"Embedding for version {version_id} already exists")
+            return True
 
-        # Build searchable text for embedding
-        searchable_text = build_searchable_text(
-            name, description, sub_heading, categories
-        )
+    # Build searchable text for embedding
+    searchable_text = build_searchable_text(name, description, sub_heading, categories)
 
-        # Generate new embedding
-        embedding = await generate_embedding(searchable_text)
-        if embedding is None:
-            logger.warning(f"Could not generate embedding for version {version_id}")
-            return False
+    # Generate new embedding
+    embedding = await generate_embedding(searchable_text)
 
-        # Store the embedding with metadata using new function
-        metadata = {
-            "name": name,
-            "subHeading": sub_heading,
-            "categories": categories,
-        }
-        return await store_content_embedding(
-            content_type=ContentType.STORE_AGENT,
-            content_id=version_id,
-            embedding=embedding,
-            searchable_text=searchable_text,
-            metadata=metadata,
-            user_id=None,  # Store agents are public
-            tx=tx,
-        )
-
-    except Exception as e:
-        logger.error(f"Failed to ensure embedding for version {version_id}: {e}")
-        return False
+    # Store the embedding with metadata using new function
+    metadata = {
+        "name": name,
+        "subHeading": sub_heading,
+        "categories": categories,
+    }
+    return await store_content_embedding(
+        content_type=ContentType.STORE_AGENT,
+        content_id=version_id,
+        embedding=embedding,
+        searchable_text=searchable_text,
+        metadata=metadata,
+        user_id=None,  # Store agents are public
+        tx=tx,
+    )
 
 
 async def delete_embedding(version_id: str) -> bool:
@@ -476,6 +454,7 @@ async def backfill_all_content_types(batch_size: int = 10) -> dict[str, Any]:
     total_processed = 0
     total_success = 0
     total_failed = 0
+    all_errors: dict[str, int] = {}  # Aggregate errors across all content types
 
     # Process content types in explicit order
     processing_order = [
@@ -521,6 +500,13 @@ async def backfill_all_content_types(batch_size: int = 10) -> dict[str, Any]:
             success = sum(1 for result in results if result is True)
             failed = len(results) - success
 
+            # Aggregate errors across all content types
+            if failed > 0:
+                for result in results:
+                    if isinstance(result, Exception):
+                        error_key = f"{type(result).__name__}: {str(result)}"
+                        all_errors[error_key] = all_errors.get(error_key, 0) + 1
+
             results_by_type[content_type.value] = {
                 "processed": len(missing_items),
                 "success": success,
@@ -546,6 +532,13 @@ async def backfill_all_content_types(batch_size: int = 10) -> dict[str, Any]:
                 "error": str(e),
             }
 
+    # Log aggregated errors once at the end
+    if all_errors:
+        error_details = ", ".join(
+            f"{error} ({count}x)" for error, count in all_errors.items()
+        )
+        logger.error(f"Embedding backfill errors: {error_details}")
+
     return {
         "by_type": results_by_type,
         "totals": {
@@ -557,11 +550,12 @@ async def backfill_all_content_types(batch_size: int = 10) -> dict[str, Any]:
     }
 
 
-async def embed_query(query: str) -> list[float] | None:
+async def embed_query(query: str) -> list[float]:
     """
     Generate embedding for a search query.
 
     Same as generate_embedding but with clearer intent.
+    Raises exceptions on failure - caller should handle.
     """
     return await generate_embedding(query)
 
@@ -594,40 +588,30 @@ async def ensure_content_embedding(
         tx: Optional transaction client
 
     Returns:
-        True if embedding exists/was created, False on failure
+        True if embedding exists/was created
+
+    Raises exceptions on failure - caller should handle.
     """
-    try:
-        # Check if embedding already exists
-        if not force:
-            existing = await get_content_embedding(content_type, content_id, user_id)
-            if existing and existing.get("embedding"):
-                logger.debug(
-                    f"Embedding for {content_type}:{content_id} already exists"
-                )
-                return True
+    # Check if embedding already exists
+    if not force:
+        existing = await get_content_embedding(content_type, content_id, user_id)
+        if existing and existing.get("embedding"):
+            logger.debug(f"Embedding for {content_type}:{content_id} already exists")
+            return True
 
-        # Generate new embedding
-        embedding = await generate_embedding(searchable_text)
-        if embedding is None:
-            logger.warning(
-                f"Could not generate embedding for {content_type}:{content_id}"
-            )
-            return False
+    # Generate new embedding
+    embedding = await generate_embedding(searchable_text)
 
-        # Store the embedding
-        return await store_content_embedding(
-            content_type=content_type,
-            content_id=content_id,
-            embedding=embedding,
-            searchable_text=searchable_text,
-            metadata=metadata or {},
-            user_id=user_id,
-            tx=tx,
-        )
-
-    except Exception as e:
-        logger.error(f"Failed to ensure embedding for {content_type}:{content_id}: {e}")
-        return False
+    # Store the embedding
+    return await store_content_embedding(
+        content_type=content_type,
+        content_id=content_id,
+        embedding=embedding,
+        searchable_text=searchable_text,
+        metadata=metadata or {},
+        user_id=user_id,
+        tx=tx,
+    )
 
 
 async def cleanup_orphaned_embeddings() -> dict[str, Any]:
@@ -854,9 +838,8 @@ async def semantic_search(
         limit = 100
 
     # Generate query embedding
-    query_embedding = await embed_query(query)
-
-    if query_embedding is not None:
+    try:
+        query_embedding = await embed_query(query)
         # Semantic search with embeddings
         embedding_str = embedding_to_vector_string(query_embedding)
 
@@ -907,24 +890,21 @@ async def semantic_search(
         """
         )
 
-        try:
-            results = await query_raw_with_schema(sql, *params)
-            return [
-                {
-                    "content_id": row["content_id"],
-                    "content_type": row["content_type"],
-                    "searchable_text": row["searchable_text"],
-                    "metadata": row["metadata"],
-                    "similarity": float(row["similarity"]),
-                }
-                for row in results
-            ]
-        except Exception as e:
-            logger.error(f"Semantic search failed: {e}")
-            # Fall through to lexical search below
+        results = await query_raw_with_schema(sql, *params)
+        return [
+            {
+                "content_id": row["content_id"],
+                "content_type": row["content_type"],
+                "searchable_text": row["searchable_text"],
+                "metadata": row["metadata"],
+                "similarity": float(row["similarity"]),
+            }
+            for row in results
+        ]
+    except Exception as e:
+        logger.warning(f"Semantic search failed, falling back to lexical search: {e}")
 
     # Fallback to lexical search if embeddings unavailable
-    logger.warning("Falling back to lexical search (embeddings unavailable)")
 
     params_lexical: list[Any] = [limit]
     user_filter = ""

autogpt_platform/backend/backend/api/features/store/embeddings_e2e_test.py

@@ -454,6 +454,9 @@ async def test_unified_hybrid_search_pagination(
     cleanup_embeddings: list,
 ):
     """Test unified search pagination works correctly."""
+    # Use a unique search term to avoid matching other test data
+    unique_term = f"xyzpagtest{uuid.uuid4().hex[:8]}"
+
     # Create multiple items
     content_ids = []
     for i in range(5):
@@ -465,14 +468,14 @@ async def test_unified_hybrid_search_pagination(
             content_type=ContentType.BLOCK,
             content_id=content_id,
             embedding=mock_embedding,
-            searchable_text=f"pagination test item number {i}",
+            searchable_text=f"{unique_term} item number {i}",
             metadata={"index": i},
             user_id=None,
         )
 
     # Get first page
     page1_results, total1 = await unified_hybrid_search(
-        query="pagination test",
+        query=unique_term,
         content_types=[ContentType.BLOCK],
         page=1,
         page_size=2,
@@ -480,7 +483,7 @@ async def test_unified_hybrid_search_pagination(
 
     # Get second page
     page2_results, total2 = await unified_hybrid_search(
-        query="pagination test",
+        query=unique_term,
         content_types=[ContentType.BLOCK],
         page=2,
         page_size=2,

autogpt_platform/backend/backend/api/features/store/embeddings_schema_test.py

@@ -298,17 +298,16 @@ async def test_schema_handling_error_cases():
             mock_client.execute_raw.side_effect = Exception("Database error")
             mock_get_client.return_value = mock_client
 
-            result = await embeddings.store_content_embedding(
-                content_type=ContentType.STORE_AGENT,
-                content_id="test-id",
-                embedding=[0.1] * EMBEDDING_DIM,
-                searchable_text="test",
-                metadata=None,
-                user_id=None,
-            )
-
-            # Should return False on error, not raise
-            assert result is False
+            # Should raise exception on error
+            with pytest.raises(Exception, match="Database error"):
+                await embeddings.store_content_embedding(
+                    content_type=ContentType.STORE_AGENT,
+                    content_id="test-id",
+                    embedding=[0.1] * EMBEDDING_DIM,
+                    searchable_text="test",
+                    metadata=None,
+                    user_id=None,
+                )
 
 
 if __name__ == "__main__":

autogpt_platform/backend/backend/api/features/store/embeddings_test.py

@@ -80,9 +80,8 @@ async def test_generate_embedding_no_api_key():
     ) as mock_get_client:
         mock_get_client.return_value = None
 
-        result = await embeddings.generate_embedding("test text")
-
-        assert result is None
+        with pytest.raises(RuntimeError, match="openai_internal_api_key not set"):
+            await embeddings.generate_embedding("test text")
 
 
 @pytest.mark.asyncio(loop_scope="session")
@@ -97,9 +96,8 @@ async def test_generate_embedding_api_error():
     ) as mock_get_client:
         mock_get_client.return_value = mock_client
 
-        result = await embeddings.generate_embedding("test text")
-
-        assert result is None
+        with pytest.raises(Exception, match="API Error"):
+            await embeddings.generate_embedding("test text")
 
 
 @pytest.mark.asyncio(loop_scope="session")
@@ -173,11 +171,10 @@ async def test_store_embedding_database_error(mocker):
 
     embedding = [0.1, 0.2, 0.3]
 
-    result = await embeddings.store_embedding(
-        version_id="test-version-id", embedding=embedding, tx=mock_client
-    )
-
-    assert result is False
+    with pytest.raises(Exception, match="Database error"):
+        await embeddings.store_embedding(
+            version_id="test-version-id", embedding=embedding, tx=mock_client
+        )
 
 
 @pytest.mark.asyncio(loop_scope="session")
@@ -277,17 +274,16 @@ async def test_ensure_embedding_create_new(mock_get, mock_store, mock_generate):
 async def test_ensure_embedding_generation_fails(mock_get, mock_generate):
     """Test ensure_embedding when generation fails."""
     mock_get.return_value = None
-    mock_generate.return_value = None
+    mock_generate.side_effect = Exception("Generation failed")
 
-    result = await embeddings.ensure_embedding(
-        version_id="test-id",
-        name="Test",
-        description="Test description",
-        sub_heading="Test heading",
-        categories=["test"],
-    )
-
-    assert result is False
+    with pytest.raises(Exception, match="Generation failed"):
+        await embeddings.ensure_embedding(
+            version_id="test-id",
+            name="Test",
+            description="Test description",
+            sub_heading="Test heading",
+            categories=["test"],
+        )
 
 
 @pytest.mark.asyncio(loop_scope="session")

autogpt_platform/backend/backend/api/features/store/hybrid_search.py

@@ -8,6 +8,7 @@ Includes BM25 reranking for improved lexical relevance.
 
 import logging
 import re
+import time
 from dataclasses import dataclass
 from typing import Any, Literal
 
@@ -186,13 +187,12 @@ async def unified_hybrid_search(
 
     offset = (page - 1) * page_size
 
-    # Generate query embedding
-    query_embedding = await embed_query(query)
-
-    # Graceful degradation if embedding unavailable
-    if query_embedding is None or not query_embedding:
+    # Generate query embedding with graceful degradation
+    try:
+        query_embedding = await embed_query(query)
+    except Exception as e:
         logger.warning(
-            "Failed to generate query embedding - falling back to lexical-only search. "
+            f"Failed to generate query embedding - falling back to lexical-only search: {e}. "
             "Check that openai_internal_api_key is configured and OpenAI API is accessible."
         )
         query_embedding = [0.0] * EMBEDDING_DIM
@@ -363,7 +363,11 @@ async def unified_hybrid_search(
         LIMIT {limit_param} OFFSET {offset_param}
     """
 
-    results = await query_raw_with_schema(sql_query, *params)
+    try:
+        results = await query_raw_with_schema(sql_query, *params)
+    except Exception as e:
+        await _log_vector_error_diagnostics(e)
+        raise
 
     total = results[0]["total_count"] if results else 0
     # Apply BM25 reranking
@@ -464,13 +468,12 @@ async def hybrid_search(
 
     offset = (page - 1) * page_size
 
-    # Generate query embedding
-    query_embedding = await embed_query(query)
-
-    # Graceful degradation
-    if query_embedding is None or not query_embedding:
+    # Generate query embedding with graceful degradation
+    try:
+        query_embedding = await embed_query(query)
+    except Exception as e:
         logger.warning(
-            "Failed to generate query embedding - falling back to lexical-only search."
+            f"Failed to generate query embedding - falling back to lexical-only search: {e}"
         )
         query_embedding = [0.0] * EMBEDDING_DIM
         total_non_semantic = (
@@ -602,6 +605,7 @@ async def hybrid_search(
                 sa.featured,
                 sa.is_available,
                 sa.updated_at,
+                sa."agentGraphId",
                 -- Searchable text for BM25 reranking
                 COALESCE(sa.agent_name, '') || ' ' || COALESCE(sa.sub_heading, '') || ' ' || COALESCE(sa.description, '') as searchable_text,
                 -- Semantic score
@@ -661,6 +665,7 @@ async def hybrid_search(
                 featured,
                 is_available,
                 updated_at,
+                "agentGraphId",
                 searchable_text,
                 semantic_score,
                 lexical_score,
@@ -686,7 +691,11 @@ async def hybrid_search(
         LIMIT {limit_param} OFFSET {offset_param}
     """
 
-    results = await query_raw_with_schema(sql_query, *params)
+    try:
+        results = await query_raw_with_schema(sql_query, *params)
+    except Exception as e:
+        await _log_vector_error_diagnostics(e)
+        raise
 
     total = results[0]["total_count"] if results else 0
 
@@ -718,6 +727,87 @@ async def hybrid_search_simple(
     return await hybrid_search(query=query, page=page, page_size=page_size)
 
 
+# ============================================================================
+# Diagnostics
+# ============================================================================
+
+# Rate limit: only log vector error diagnostics once per this interval
+_VECTOR_DIAG_INTERVAL_SECONDS = 60
+_last_vector_diag_time: float = 0
+
+
+async def _log_vector_error_diagnostics(error: Exception) -> None:
+    """Log diagnostic info when 'type vector does not exist' error occurs.
+
+    Note: Diagnostic queries use query_raw_with_schema which may run on a different
+    pooled connection than the one that failed. Session-level search_path can differ,
+    so these diagnostics show cluster-wide state, not necessarily the failed session.
+
+    Includes rate limiting to avoid log spam - only logs once per minute.
+    Caller should re-raise the error after calling this function.
+    """
+    global _last_vector_diag_time
+
+    # Check if this is the vector type error
+    error_str = str(error).lower()
+    if not (
+        "type" in error_str and "vector" in error_str and "does not exist" in error_str
+    ):
+        return
+
+    # Rate limit: only log once per interval
+    now = time.time()
+    if now - _last_vector_diag_time < _VECTOR_DIAG_INTERVAL_SECONDS:
+        return
+    _last_vector_diag_time = now
+
+    try:
+        diagnostics: dict[str, object] = {}
+
+        try:
+            search_path_result = await query_raw_with_schema("SHOW search_path")
+            diagnostics["search_path"] = search_path_result
+        except Exception as e:
+            diagnostics["search_path"] = f"Error: {e}"
+
+        try:
+            schema_result = await query_raw_with_schema("SELECT current_schema()")
+            diagnostics["current_schema"] = schema_result
+        except Exception as e:
+            diagnostics["current_schema"] = f"Error: {e}"
+
+        try:
+            user_result = await query_raw_with_schema(
+                "SELECT current_user, session_user, current_database()"
+            )
+            diagnostics["user_info"] = user_result
+        except Exception as e:
+            diagnostics["user_info"] = f"Error: {e}"
+
+        try:
+            # Check pgvector extension installation (cluster-wide, stable info)
+            ext_result = await query_raw_with_schema(
+                "SELECT extname, extversion, nspname as schema "
+                "FROM pg_extension e "
+                "JOIN pg_namespace n ON e.extnamespace = n.oid "
+                "WHERE extname = 'vector'"
+            )
+            diagnostics["pgvector_extension"] = ext_result
+        except Exception as e:
+            diagnostics["pgvector_extension"] = f"Error: {e}"
+
+        logger.error(
+            f"Vector type error diagnostics:\n"
+            f"  Error: {error}\n"
+            f"  search_path: {diagnostics.get('search_path')}\n"
+            f"  current_schema: {diagnostics.get('current_schema')}\n"
+            f"  user_info: {diagnostics.get('user_info')}\n"
+            f"  pgvector_extension: {diagnostics.get('pgvector_extension')}"
+        )
+    except Exception as diag_error:
+        logger.error(f"Failed to collect vector error diagnostics: {diag_error}")
+
+
 # Backward compatibility alias - HybridSearchWeights maps to StoreAgentSearchWeights
 # for existing code that expects the popularity parameter
 HybridSearchWeights = StoreAgentSearchWeights

autogpt_platform/backend/backend/api/features/store/hybrid_search_test.py

@@ -172,8 +172,8 @@ async def test_hybrid_search_without_embeddings():
         with patch(
             "backend.api.features.store.hybrid_search.query_raw_with_schema"
         ) as mock_query:
-            # Simulate embedding failure
-            mock_embed.return_value = None
+            # Simulate embedding failure by raising exception
+            mock_embed.side_effect = Exception("Embedding generation failed")
             mock_query.return_value = mock_results
 
             # Should NOT raise - graceful degradation
@@ -613,7 +613,9 @@ async def test_unified_hybrid_search_graceful_degradation():
             "backend.api.features.store.hybrid_search.embed_query"
         ) as mock_embed:
             mock_query.return_value = mock_results
-            mock_embed.return_value = None  # Embedding failure
+            mock_embed.side_effect = Exception(
+                "Embedding generation failed"
+            )  # Embedding failure
 
             # Should NOT raise - graceful degradation
             results, total = await unified_hybrid_search(

autogpt_platform/backend/backend/api/features/store/image_gen.py

@@ -16,7 +16,7 @@ from backend.blocks.ideogram import (
     StyleType,
     UpscaleOption,
 )
-from backend.data.graph import BaseGraph
+from backend.data.graph import GraphBaseMeta
 from backend.data.model import CredentialsMetaInput, ProviderName
 from backend.integrations.credentials_store import ideogram_credentials
 from backend.util.request import Requests
@@ -34,14 +34,14 @@ class ImageStyle(str, Enum):
     DIGITAL_ART = "digital art"
 
 
-async def generate_agent_image(agent: BaseGraph | AgentGraph) -> io.BytesIO:
+async def generate_agent_image(agent: GraphBaseMeta | AgentGraph) -> io.BytesIO:
     if settings.config.use_agent_image_generation_v2:
         return await generate_agent_image_v2(graph=agent)
     else:
         return await generate_agent_image_v1(agent=agent)
 
 
-async def generate_agent_image_v2(graph: BaseGraph | AgentGraph) -> io.BytesIO:
+async def generate_agent_image_v2(graph: GraphBaseMeta | AgentGraph) -> io.BytesIO:
     """
     Generate an image for an agent using Ideogram model.
     Returns:
@@ -54,14 +54,17 @@ async def generate_agent_image_v2(graph: BaseGraph | AgentGraph) -> io.BytesIO:
     description = f"{name} ({graph.description})" if graph.description else name
 
     prompt = (
-        f"Create a visually striking retro-futuristic vector pop art illustration prominently featuring "
-        f'"{name}" in bold typography. The image clearly and literally depicts a {description}, '
-        f"along with recognizable objects directly associated with the primary function of a {name}. "
-        f"Ensure the imagery is concrete, intuitive, and immediately understandable, clearly conveying the "
-        f"purpose of a {name}. Maintain vibrant, limited-palette colors, sharp vector lines, geometric "
-        f"shapes, flat illustration techniques, and solid colors without gradients or shading. Preserve a "
-        f"retro-futuristic aesthetic influenced by mid-century futurism and 1960s psychedelia, "
-        f"prioritizing clear visual storytelling and thematic clarity above all else."
+        "Create a visually striking retro-futuristic vector pop art illustration "
+        f'prominently featuring "{name}" in bold typography. The image clearly and '
+        f"literally depicts a {description}, along with recognizable objects directly "
+        f"associated with the primary function of a {name}. "
+        f"Ensure the imagery is concrete, intuitive, and immediately understandable, "
+        f"clearly conveying the purpose of a {name}. "
+        "Maintain vibrant, limited-palette colors, sharp vector lines, "
+        "geometric shapes, flat illustration techniques, and solid colors "
+        "without gradients or shading. Preserve a retro-futuristic aesthetic "
+        "influenced by mid-century futurism and 1960s psychedelia, "
+        "prioritizing clear visual storytelling and thematic clarity above all else."
     )
 
     custom_colors = [
@@ -99,12 +102,12 @@ async def generate_agent_image_v2(graph: BaseGraph | AgentGraph) -> io.BytesIO:
     return io.BytesIO(response.content)
 
 
-async def generate_agent_image_v1(agent: BaseGraph | AgentGraph) -> io.BytesIO:
+async def generate_agent_image_v1(agent: GraphBaseMeta | AgentGraph) -> io.BytesIO:
     """
     Generate an image for an agent using Flux model via Replicate API.
 
     Args:
-        agent (Graph): The agent to generate an image for
+        agent (GraphBaseMeta | AgentGraph): The agent to generate an image for
 
     Returns:
         io.BytesIO: The generated image as bytes
@@ -114,7 +117,13 @@ async def generate_agent_image_v1(agent: BaseGraph | AgentGraph) -> io.BytesIO:
             raise ValueError("Missing Replicate API key in settings")
 
         # Construct prompt from agent details
-        prompt = f"Create a visually engaging app store thumbnail for the AI agent that highlights what it does in a clear and captivating way:\n- **Name**: {agent.name}\n- **Description**: {agent.description}\nFocus on showcasing its core functionality with an appealing design."
+        prompt = (
+            "Create a visually engaging app store thumbnail for the AI agent "
+            "that highlights what it does in a clear and captivating way:\n"
+            f"- **Name**: {agent.name}\n"
+            f"- **Description**: {agent.description}\n"
+            f"Focus on showcasing its core functionality with an appealing design."
+        )
 
         # Set up Replicate client
         client = ReplicateClient(api_token=settings.secrets.replicate_api_key)

autogpt_platform/backend/backend/api/features/store/model.py

@@ -38,6 +38,7 @@ class StoreAgent(pydantic.BaseModel):
     description: str
     runs: int
     rating: float
+    agent_graph_id: str
 
 
 class StoreAgentsResponse(pydantic.BaseModel):

autogpt_platform/backend/backend/api/features/store/model_test.py

@@ -26,11 +26,13 @@ def test_store_agent():
         description="Test description",
         runs=50,
         rating=4.5,
+        agent_graph_id="test-graph-id",
     )
     assert agent.slug == "test-agent"
     assert agent.agent_name == "Test Agent"
     assert agent.runs == 50
     assert agent.rating == 4.5
+    assert agent.agent_graph_id == "test-graph-id"
 
 
 def test_store_agents_response():
@@ -46,6 +48,7 @@ def test_store_agents_response():
                 description="Test description",
                 runs=50,
                 rating=4.5,
+                agent_graph_id="test-graph-id",
             )
         ],
         pagination=store_model.Pagination(

autogpt_platform/backend/backend/api/features/store/routes.py

@@ -278,7 +278,7 @@ async def get_agent(
 )
 async def get_graph_meta_by_store_listing_version_id(
     store_listing_version_id: str,
-) -> backend.data.graph.GraphMeta:
+) -> backend.data.graph.GraphModelWithoutNodes:
     """
     Get Agent Graph from Store Listing Version ID.
     """

autogpt_platform/backend/backend/api/features/store/routes_test.py

@@ -82,6 +82,7 @@ def test_get_agents_featured(
                 description="Featured agent description",
                 runs=100,
                 rating=4.5,
+                agent_graph_id="test-graph-1",
             )
         ],
         pagination=store_model.Pagination(
@@ -127,6 +128,7 @@ def test_get_agents_by_creator(
                 description="Creator agent description",
                 runs=50,
                 rating=4.0,
+                agent_graph_id="test-graph-2",
             )
         ],
         pagination=store_model.Pagination(
@@ -172,6 +174,7 @@ def test_get_agents_sorted(
                 description="Top agent description",
                 runs=1000,
                 rating=5.0,
+                agent_graph_id="test-graph-3",
             )
         ],
         pagination=store_model.Pagination(
@@ -217,6 +220,7 @@ def test_get_agents_search(
                 description="Specific search term description",
                 runs=75,
                 rating=4.2,
+                agent_graph_id="test-graph-search",
             )
         ],
         pagination=store_model.Pagination(
@@ -262,6 +266,7 @@ def test_get_agents_category(
                 description="Category agent description",
                 runs=60,
                 rating=4.1,
+                agent_graph_id="test-graph-category",
             )
         ],
         pagination=store_model.Pagination(
@@ -306,6 +311,7 @@ def test_get_agents_pagination(
                 description=f"Agent {i} description",
                 runs=i * 10,
                 rating=4.0,
+                agent_graph_id="test-graph-2",
             )
             for i in range(5)
         ],

autogpt_platform/backend/backend/api/features/store/test_cache_delete.py

@@ -33,6 +33,7 @@ class TestCacheDeletion:
                     description="Test description",
                     runs=100,
                     rating=4.5,
+                    agent_graph_id="test-graph-id",
                 )
             ],
             pagination=Pagination(

autogpt_platform/backend/backend/api/features/v1.py

@@ -101,7 +101,6 @@ from backend.util.timezone_utils import (
 from backend.util.virus_scanner import scan_content_safe
 
 from .library import db as library_db
-from .library import model as library_model
 from .store.model import StoreAgentDetails
 
 
@@ -261,14 +260,36 @@ async def get_onboarding_agents(
     return await get_recommended_agents(user_id)
 
 
+class OnboardingStatusResponse(pydantic.BaseModel):
+    """Response for onboarding status check."""
+
+    is_onboarding_enabled: bool
+    is_chat_enabled: bool
+
+
 @v1_router.get(
     "/onboarding/enabled",
     summary="Is onboarding enabled",
     tags=["onboarding", "public"],
-    dependencies=[Security(requires_user)],
+    response_model=OnboardingStatusResponse,
 )
-async def is_onboarding_enabled() -> bool:
-    return await onboarding_enabled()
+async def is_onboarding_enabled(
+    user_id: Annotated[str, Security(get_user_id)],
+) -> OnboardingStatusResponse:
+    # Check if chat is enabled for user
+    is_chat_enabled = await is_feature_enabled(Flag.CHAT, user_id, False)
+
+    # If chat is enabled, skip legacy onboarding
+    if is_chat_enabled:
+        return OnboardingStatusResponse(
+            is_onboarding_enabled=False,
+            is_chat_enabled=True,
+        )
+
+    return OnboardingStatusResponse(
+        is_onboarding_enabled=await onboarding_enabled(),
+        is_chat_enabled=False,
+    )
 
 
 @v1_router.post(
@@ -364,6 +385,8 @@ async def execute_graph_block(
     obj = get_block(block_id)
     if not obj:
         raise HTTPException(status_code=404, detail=f"Block #{block_id} not found.")
+    if obj.disabled:
+        raise HTTPException(status_code=403, detail=f"Block #{block_id} is disabled.")
 
     user = await get_user_by_id(user_id)
     if not user:
@@ -799,18 +822,16 @@ async def update_graph(
     graph: graph_db.Graph,
     user_id: Annotated[str, Security(get_user_id)],
 ) -> graph_db.GraphModel:
-    # Sanity check
     if graph.id and graph.id != graph_id:
         raise HTTPException(400, detail="Graph ID does not match ID in URI")
 
-    # Determine new version
     existing_versions = await graph_db.get_graph_all_versions(graph_id, user_id=user_id)
     if not existing_versions:
         raise HTTPException(404, detail=f"Graph #{graph_id} not found")
-    latest_version_number = max(g.version for g in existing_versions)
-    graph.version = latest_version_number + 1
 
+    graph.version = max(g.version for g in existing_versions) + 1
     current_active_version = next((v for v in existing_versions if v.is_active), None)
+
     graph = graph_db.make_graph_model(graph, user_id)
     graph.reassign_ids(user_id=user_id, reassign_graph_id=False)
     graph.validate_graph(for_run=False)
@@ -818,27 +839,23 @@ async def update_graph(
     new_graph_version = await graph_db.create_graph(graph, user_id=user_id)
 
     if new_graph_version.is_active:
-        # Keep the library agent up to date with the new active version
-        await _update_library_agent_version_and_settings(user_id, new_graph_version)
-
-        # Handle activation of the new graph first to ensure continuity
+        await library_db.update_library_agent_version_and_settings(
+            user_id, new_graph_version
+        )
         new_graph_version = await on_graph_activate(new_graph_version, user_id=user_id)
-        # Ensure new version is the only active version
         await graph_db.set_graph_active_version(
             graph_id=graph_id, version=new_graph_version.version, user_id=user_id
         )
         if current_active_version:
-            # Handle deactivation of the previously active version
             await on_graph_deactivate(current_active_version, user_id=user_id)
 
-    # Fetch new graph version *with sub-graphs* (needed for credentials input schema)
     new_graph_version_with_subgraphs = await graph_db.get_graph(
         graph_id,
         new_graph_version.version,
         user_id=user_id,
         include_subgraphs=True,
     )
-    assert new_graph_version_with_subgraphs  # make type checker happy
+    assert new_graph_version_with_subgraphs
     return new_graph_version_with_subgraphs
 
 
@@ -876,33 +893,15 @@ async def set_graph_active_version(
     )
 
     # Keep the library agent up to date with the new active version
-    await _update_library_agent_version_and_settings(user_id, new_active_graph)
+    await library_db.update_library_agent_version_and_settings(
+        user_id, new_active_graph
+    )
 
     if current_active_graph and current_active_graph.version != new_active_version:
         # Handle deactivation of the previously active version
         await on_graph_deactivate(current_active_graph, user_id=user_id)
 
 
-async def _update_library_agent_version_and_settings(
-    user_id: str, agent_graph: graph_db.GraphModel
-) -> library_model.LibraryAgent:
-    library = await library_db.update_agent_version_in_library(
-        user_id, agent_graph.id, agent_graph.version
-    )
-    updated_settings = GraphSettings.from_graph(
-        graph=agent_graph,
-        hitl_safe_mode=library.settings.human_in_the_loop_safe_mode,
-        sensitive_action_safe_mode=library.settings.sensitive_action_safe_mode,
-    )
-    if updated_settings != library.settings:
-        library = await library_db.update_library_agent(
-            library_agent_id=library.id,
-            user_id=user_id,
-            settings=updated_settings,
-        )
-    return library
-
-
 @v1_router.patch(
     path="/graphs/{graph_id}/settings",
     summary="Update graph settings",

autogpt_platform/backend/backend/api/features/v1_test.py

@@ -138,6 +138,7 @@ def test_execute_graph_block(
     """Test execute block endpoint"""
     # Mock block
     mock_block = Mock()
+    mock_block.disabled = False
 
     async def mock_execute(*args, **kwargs):
         yield "output1", {"data": "result1"}

autogpt_platform/backend/backend/api/features/workspace/__init__.py

@@ -0,0 +1 @@
+# Workspace API feature module

autogpt_platform/backend/backend/api/features/workspace/routes.py

@@ -0,0 +1,122 @@
+"""
+Workspace API routes for managing user file storage.
+"""
+
+import logging
+import re
+from typing import Annotated
+from urllib.parse import quote
+
+import fastapi
+from autogpt_libs.auth.dependencies import get_user_id, requires_user
+from fastapi.responses import Response
+
+from backend.data.workspace import get_workspace, get_workspace_file
+from backend.util.workspace_storage import get_workspace_storage
+
+
+def _sanitize_filename_for_header(filename: str) -> str:
+    """
+    Sanitize filename for Content-Disposition header to prevent header injection.
+
+    Removes/replaces characters that could break the header or inject new headers.
+    Uses RFC5987 encoding for non-ASCII characters.
+    """
+    # Remove CR, LF, and null bytes (header injection prevention)
+    sanitized = re.sub(r"[\r\n\x00]", "", filename)
+    # Escape quotes
+    sanitized = sanitized.replace('"', '\\"')
+    # For non-ASCII, use RFC5987 filename* parameter
+    # Check if filename has non-ASCII characters
+    try:
+        sanitized.encode("ascii")
+        return f'attachment; filename="{sanitized}"'
+    except UnicodeEncodeError:
+        # Use RFC5987 encoding for UTF-8 filenames
+        encoded = quote(sanitized, safe="")
+        return f"attachment; filename*=UTF-8''{encoded}"
+
+
+logger = logging.getLogger(__name__)
+
+router = fastapi.APIRouter(
+    dependencies=[fastapi.Security(requires_user)],
+)
+
+
+def _create_streaming_response(content: bytes, file) -> Response:
+    """Create a streaming response for file content."""
+    return Response(
+        content=content,
+        media_type=file.mimeType,
+        headers={
+            "Content-Disposition": _sanitize_filename_for_header(file.name),
+            "Content-Length": str(len(content)),
+        },
+    )
+
+
+async def _create_file_download_response(file) -> Response:
+    """
+    Create a download response for a workspace file.
+
+    Handles both local storage (direct streaming) and GCS (signed URL redirect
+    with fallback to streaming).
+    """
+    storage = await get_workspace_storage()
+
+    # For local storage, stream the file directly
+    if file.storagePath.startswith("local://"):
+        content = await storage.retrieve(file.storagePath)
+        return _create_streaming_response(content, file)
+
+    # For GCS, try to redirect to signed URL, fall back to streaming
+    try:
+        url = await storage.get_download_url(file.storagePath, expires_in=300)
+        # If we got back an API path (fallback), stream directly instead
+        if url.startswith("/api/"):
+            content = await storage.retrieve(file.storagePath)
+            return _create_streaming_response(content, file)
+        return fastapi.responses.RedirectResponse(url=url, status_code=302)
+    except Exception as e:
+        # Log the signed URL failure with context
+        logger.error(
+            f"Failed to get signed URL for file {file.id} "
+            f"(storagePath={file.storagePath}): {e}",
+            exc_info=True,
+        )
+        # Fall back to streaming directly from GCS
+        try:
+            content = await storage.retrieve(file.storagePath)
+            return _create_streaming_response(content, file)
+        except Exception as fallback_error:
+            logger.error(
+                f"Fallback streaming also failed for file {file.id} "
+                f"(storagePath={file.storagePath}): {fallback_error}",
+                exc_info=True,
+            )
+            raise
+
+
+@router.get(
+    "/files/{file_id}/download",
+    summary="Download file by ID",
+)
+async def download_file(
+    user_id: Annotated[str, fastapi.Security(get_user_id)],
+    file_id: str,
+) -> Response:
+    """
+    Download a file by its ID.
+
+    Returns the file content directly or redirects to a signed URL for GCS.
+    """
+    workspace = await get_workspace(user_id)
+    if workspace is None:
+        raise fastapi.HTTPException(status_code=404, detail="Workspace not found")
+
+    file = await get_workspace_file(file_id, workspace.id)
+    if file is None:
+        raise fastapi.HTTPException(status_code=404, detail="File not found")
+
+    return await _create_file_download_response(file)

autogpt_platform/backend/backend/api/rest_api.py

@@ -32,6 +32,7 @@ import backend.api.features.postmark.postmark
 import backend.api.features.store.model
 import backend.api.features.store.routes
 import backend.api.features.v1
+import backend.api.features.workspace.routes as workspace_routes
 import backend.data.block
 import backend.data.db
 import backend.data.graph
@@ -40,6 +41,10 @@ import backend.integrations.webhooks.utils
 import backend.util.service
 import backend.util.settings
 from backend.blocks.llm import DEFAULT_LLM_MODEL
+from backend.copilot.completion_consumer import (
+    start_completion_consumer,
+    stop_completion_consumer,
+)
 from backend.data.model import Credentials
 from backend.integrations.providers import ProviderName
 from backend.monitoring.instrumentation import instrument_fastapi
@@ -52,6 +57,7 @@ from backend.util.exceptions import (
 )
 from backend.util.feature_flag import initialize_launchdarkly, shutdown_launchdarkly
 from backend.util.service import UnhealthyServiceError
+from backend.util.workspace_storage import shutdown_workspace_storage
 
 from .external.fastapi_app import external_api
 from .features.analytics import router as analytics_router
@@ -116,14 +122,31 @@ async def lifespan_context(app: fastapi.FastAPI):
     await backend.data.graph.migrate_llm_models(DEFAULT_LLM_MODEL)
     await backend.integrations.webhooks.utils.migrate_legacy_triggered_graphs()
 
+    # Start chat completion consumer for Redis Streams notifications
+    try:
+        await start_completion_consumer()
+    except Exception as e:
+        logger.warning(f"Could not start chat completion consumer: {e}")
+
     with launch_darkly_context():
         yield
 
+    # Stop chat completion consumer
+    try:
+        await stop_completion_consumer()
+    except Exception as e:
+        logger.warning(f"Error stopping chat completion consumer: {e}")
+
     try:
         await shutdown_cloud_storage_handler()
     except Exception as e:
         logger.warning(f"Error shutting down cloud storage handler: {e}")
 
+    try:
+        await shutdown_workspace_storage()
+    except Exception as e:
+        logger.warning(f"Error shutting down workspace storage: {e}")
+
     await backend.data.db.disconnect()
 
 
@@ -315,6 +338,11 @@ app.include_router(
     tags=["v2", "chat"],
     prefix="/api/chat",
 )
+app.include_router(
+    workspace_routes.router,
+    tags=["workspace"],
+    prefix="/api/workspace",
+)
 app.include_router(
     backend.api.features.oauth.router,
     tags=["oauth"],

autogpt_platform/backend/backend/api/ws_api.py

@@ -66,18 +66,24 @@ async def event_broadcaster(manager: ConnectionManager):
     execution_bus = AsyncRedisExecutionEventBus()
     notification_bus = AsyncRedisNotificationEventBus()
 
-    async def execution_worker():
-        async for event in execution_bus.listen("*"):
-            await manager.send_execution_update(event)
+    try:
 
-    async def notification_worker():
-        async for notification in notification_bus.listen("*"):
-            await manager.send_notification(
-                user_id=notification.user_id,
-                payload=notification.payload,
-            )
+        async def execution_worker():
+            async for event in execution_bus.listen("*"):
+                await manager.send_execution_update(event)
 
-    await asyncio.gather(execution_worker(), notification_worker())
+        async def notification_worker():
+            async for notification in notification_bus.listen("*"):
+                await manager.send_notification(
+                    user_id=notification.user_id,
+                    payload=notification.payload,
+                )
+
+        await asyncio.gather(execution_worker(), notification_worker())
+    finally:
+        # Ensure PubSub connections are closed on any exit to prevent leaks
+        await execution_bus.close()
+        await notification_bus.close()
 
 
 async def authenticate_websocket(websocket: WebSocket) -> str:

autogpt_platform/backend/backend/app.py

@@ -38,6 +38,7 @@ def main(**kwargs):
 
     from backend.api.rest_api import AgentServer
     from backend.api.ws_api import WebsocketServer
+    from backend.copilot.executor.manager import CoPilotExecutor
     from backend.executor import DatabaseManager, ExecutionManager, Scheduler
     from backend.notifications import NotificationManager
 
@@ -48,6 +49,7 @@ def main(**kwargs):
         WebsocketServer(),
         AgentServer(),
         ExecutionManager(),
+        CoPilotExecutor(),
         **kwargs,
     )
 

autogpt_platform/backend/backend/blocks/ai_image_customizer.py

@@ -13,6 +13,7 @@ from backend.data.block import (
     BlockSchemaInput,
     BlockSchemaOutput,
 )
+from backend.data.execution import ExecutionContext
 from backend.data.model import (
     APIKeyCredentials,
     CredentialsField,
@@ -117,11 +118,13 @@ class AIImageCustomizerBlock(Block):
                 "credentials": TEST_CREDENTIALS_INPUT,
             },
             test_output=[
-                ("image_url", "https://replicate.delivery/generated-image.jpg"),
+                # Output will be a workspace ref or data URI depending on context
+                ("image_url", lambda x: x.startswith(("workspace://", "data:"))),
             ],
             test_mock={
+                # Use data URI to avoid HTTP requests during tests
                 "run_model": lambda *args, **kwargs: MediaFileType(
-                    "https://replicate.delivery/generated-image.jpg"
+                    "data:image/jpeg;base64,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"
                 ),
             },
             test_credentials=TEST_CREDENTIALS,
@@ -132,8 +135,7 @@ class AIImageCustomizerBlock(Block):
         input_data: Input,
         *,
         credentials: APIKeyCredentials,
-        graph_exec_id: str,
-        user_id: str,
+        execution_context: ExecutionContext,
         **kwargs,
     ) -> BlockOutput:
         try:
@@ -141,10 +143,9 @@ class AIImageCustomizerBlock(Block):
             processed_images = await asyncio.gather(
                 *(
                     store_media_file(
-                        graph_exec_id=graph_exec_id,
                         file=img,
-                        user_id=user_id,
-                        return_content=True,
+                        execution_context=execution_context,
+                        return_format="for_external_api",  # Get content for Replicate API
                     )
                     for img in input_data.images
                 )
@@ -158,7 +159,14 @@ class AIImageCustomizerBlock(Block):
                 aspect_ratio=input_data.aspect_ratio.value,
                 output_format=input_data.output_format.value,
             )
-            yield "image_url", result
+
+            # Store the generated image to the user's workspace for persistence
+            stored_url = await store_media_file(
+                file=result,
+                execution_context=execution_context,
+                return_format="for_block_output",
+            )
+            yield "image_url", stored_url
         except Exception as e:
             yield "error", str(e)
 

autogpt_platform/backend/backend/blocks/ai_image_generator_block.py

@@ -6,6 +6,7 @@ from replicate.client import Client as ReplicateClient
 from replicate.helpers import FileOutput
 
 from backend.data.block import Block, BlockCategory, BlockSchemaInput, BlockSchemaOutput
+from backend.data.execution import ExecutionContext
 from backend.data.model import (
     APIKeyCredentials,
     CredentialsField,
@@ -13,6 +14,8 @@ from backend.data.model import (
     SchemaField,
 )
 from backend.integrations.providers import ProviderName
+from backend.util.file import store_media_file
+from backend.util.type import MediaFileType
 
 
 class ImageSize(str, Enum):
@@ -165,11 +168,13 @@ class AIImageGeneratorBlock(Block):
             test_output=[
                 (
                     "image_url",
-                    "https://replicate.delivery/generated-image.webp",
+                    # Test output is a data URI since we now store images
+                    lambda x: x.startswith("data:image/"),
                 ),
             ],
             test_mock={
-                "_run_client": lambda *args, **kwargs: "https://replicate.delivery/generated-image.webp"
+                # Return a data URI directly so store_media_file doesn't need to download
+                "_run_client": lambda *args, **kwargs: "data:image/webp;base64,UklGRiQAAABXRUJQVlA4IBgAAAAwAQCdASoBAAEAAQAcJYgCdAEO"
             },
         )
 
@@ -318,11 +323,24 @@ class AIImageGeneratorBlock(Block):
         style_text = style_map.get(style, "")
         return f"{style_text} of" if style_text else ""
 
-    async def run(self, input_data: Input, *, credentials: APIKeyCredentials, **kwargs):
+    async def run(
+        self,
+        input_data: Input,
+        *,
+        credentials: APIKeyCredentials,
+        execution_context: ExecutionContext,
+        **kwargs,
+    ):
         try:
             url = await self.generate_image(input_data, credentials)
             if url:
-                yield "image_url", url
+                # Store the generated image to the user's workspace/execution folder
+                stored_url = await store_media_file(
+                    file=MediaFileType(url),
+                    execution_context=execution_context,
+                    return_format="for_block_output",
+                )
+                yield "image_url", stored_url
             else:
                 yield "error", "Image generation returned an empty result."
         except Exception as e:

autogpt_platform/backend/backend/blocks/ai_shortform_video_block.py

@@ -13,6 +13,7 @@ from backend.data.block import (
     BlockSchemaInput,
     BlockSchemaOutput,
 )
+from backend.data.execution import ExecutionContext
 from backend.data.model import (
     APIKeyCredentials,
     CredentialsField,
@@ -21,7 +22,9 @@ from backend.data.model import (
 )
 from backend.integrations.providers import ProviderName
 from backend.util.exceptions import BlockExecutionError
+from backend.util.file import store_media_file
 from backend.util.request import Requests
+from backend.util.type import MediaFileType
 
 TEST_CREDENTIALS = APIKeyCredentials(
     id="01234567-89ab-cdef-0123-456789abcdef",
@@ -271,7 +274,10 @@ class AIShortformVideoCreatorBlock(Block):
                 "voice": Voice.LILY,
                 "video_style": VisualMediaType.STOCK_VIDEOS,
             },
-            test_output=("video_url", "https://example.com/video.mp4"),
+            test_output=(
+                "video_url",
+                lambda x: x.startswith(("workspace://", "data:")),
+            ),
             test_mock={
                 "create_webhook": lambda *args, **kwargs: (
                     "test_uuid",
@@ -280,15 +286,21 @@ class AIShortformVideoCreatorBlock(Block):
                 "create_video": lambda *args, **kwargs: {"pid": "test_pid"},
                 "check_video_status": lambda *args, **kwargs: {
                     "status": "ready",
-                    "videoUrl": "https://example.com/video.mp4",
+                    "videoUrl": "data:video/mp4;base64,AAAA",
                 },
-                "wait_for_video": lambda *args, **kwargs: "https://example.com/video.mp4",
+                # Use data URI to avoid HTTP requests during tests
+                "wait_for_video": lambda *args, **kwargs: "data:video/mp4;base64,AAAA",
             },
             test_credentials=TEST_CREDENTIALS,
         )
 
     async def run(
-        self, input_data: Input, *, credentials: APIKeyCredentials, **kwargs
+        self,
+        input_data: Input,
+        *,
+        credentials: APIKeyCredentials,
+        execution_context: ExecutionContext,
+        **kwargs,
     ) -> BlockOutput:
         # Create a new Webhook.site URL
         webhook_token, webhook_url = await self.create_webhook()
@@ -340,7 +352,13 @@ class AIShortformVideoCreatorBlock(Block):
             )
             video_url = await self.wait_for_video(credentials.api_key, pid)
             logger.debug(f"Video ready: {video_url}")
-            yield "video_url", video_url
+            # Store the generated video to the user's workspace for persistence
+            stored_url = await store_media_file(
+                file=MediaFileType(video_url),
+                execution_context=execution_context,
+                return_format="for_block_output",
+            )
+            yield "video_url", stored_url
 
 
 class AIAdMakerVideoCreatorBlock(Block):
@@ -447,7 +465,10 @@ class AIAdMakerVideoCreatorBlock(Block):
                     "https://cdn.revid.ai/uploads/1747076315114-image.png",
                 ],
             },
-            test_output=("video_url", "https://example.com/ad.mp4"),
+            test_output=(
+                "video_url",
+                lambda x: x.startswith(("workspace://", "data:")),
+            ),
             test_mock={
                 "create_webhook": lambda *args, **kwargs: (
                     "test_uuid",
@@ -456,14 +477,21 @@ class AIAdMakerVideoCreatorBlock(Block):
                 "create_video": lambda *args, **kwargs: {"pid": "test_pid"},
                 "check_video_status": lambda *args, **kwargs: {
                     "status": "ready",
-                    "videoUrl": "https://example.com/ad.mp4",
+                    "videoUrl": "data:video/mp4;base64,AAAA",
                 },
-                "wait_for_video": lambda *args, **kwargs: "https://example.com/ad.mp4",
+                "wait_for_video": lambda *args, **kwargs: "data:video/mp4;base64,AAAA",
             },
             test_credentials=TEST_CREDENTIALS,
         )
 
-    async def run(self, input_data: Input, *, credentials: APIKeyCredentials, **kwargs):
+    async def run(
+        self,
+        input_data: Input,
+        *,
+        credentials: APIKeyCredentials,
+        execution_context: ExecutionContext,
+        **kwargs,
+    ):
         webhook_token, webhook_url = await self.create_webhook()
 
         payload = {
@@ -531,7 +559,13 @@ class AIAdMakerVideoCreatorBlock(Block):
             raise RuntimeError("Failed to create video: No project ID returned")
 
         video_url = await self.wait_for_video(credentials.api_key, pid)
-        yield "video_url", video_url
+        # Store the generated video to the user's workspace for persistence
+        stored_url = await store_media_file(
+            file=MediaFileType(video_url),
+            execution_context=execution_context,
+            return_format="for_block_output",
+        )
+        yield "video_url", stored_url
 
 
 class AIScreenshotToVideoAdBlock(Block):
@@ -626,7 +660,10 @@ class AIScreenshotToVideoAdBlock(Block):
                 "script": "Amazing numbers!",
                 "screenshot_url": "https://cdn.revid.ai/uploads/1747080376028-image.png",
             },
-            test_output=("video_url", "https://example.com/screenshot.mp4"),
+            test_output=(
+                "video_url",
+                lambda x: x.startswith(("workspace://", "data:")),
+            ),
             test_mock={
                 "create_webhook": lambda *args, **kwargs: (
                     "test_uuid",
@@ -635,14 +672,21 @@ class AIScreenshotToVideoAdBlock(Block):
                 "create_video": lambda *args, **kwargs: {"pid": "test_pid"},
                 "check_video_status": lambda *args, **kwargs: {
                     "status": "ready",
-                    "videoUrl": "https://example.com/screenshot.mp4",
+                    "videoUrl": "data:video/mp4;base64,AAAA",
                 },
-                "wait_for_video": lambda *args, **kwargs: "https://example.com/screenshot.mp4",
+                "wait_for_video": lambda *args, **kwargs: "data:video/mp4;base64,AAAA",
             },
             test_credentials=TEST_CREDENTIALS,
         )
 
-    async def run(self, input_data: Input, *, credentials: APIKeyCredentials, **kwargs):
+    async def run(
+        self,
+        input_data: Input,
+        *,
+        credentials: APIKeyCredentials,
+        execution_context: ExecutionContext,
+        **kwargs,
+    ):
         webhook_token, webhook_url = await self.create_webhook()
 
         payload = {
@@ -710,4 +754,10 @@ class AIScreenshotToVideoAdBlock(Block):
             raise RuntimeError("Failed to create video: No project ID returned")
 
         video_url = await self.wait_for_video(credentials.api_key, pid)
-        yield "video_url", video_url
+        # Store the generated video to the user's workspace for persistence
+        stored_url = await store_media_file(
+            file=MediaFileType(video_url),
+            execution_context=execution_context,
+            return_format="for_block_output",
+        )
+        yield "video_url", stored_url

autogpt_platform/backend/backend/blocks/bannerbear/text_overlay.py

@@ -6,6 +6,7 @@ if TYPE_CHECKING:
 
 from pydantic import SecretStr
 
+from backend.data.execution import ExecutionContext
 from backend.sdk import (
     APIKeyCredentials,
     Block,
@@ -17,6 +18,8 @@ from backend.sdk import (
     Requests,
     SchemaField,
 )
+from backend.util.file import store_media_file
+from backend.util.type import MediaFileType
 
 from ._config import bannerbear
 
@@ -135,15 +138,17 @@ class BannerbearTextOverlayBlock(Block):
             },
             test_output=[
                 ("success", True),
-                ("image_url", "https://cdn.bannerbear.com/test-image.jpg"),
+                # Output will be a workspace ref or data URI depending on context
+                ("image_url", lambda x: x.startswith(("workspace://", "data:"))),
                 ("uid", "test-uid-123"),
                 ("status", "completed"),
             ],
             test_mock={
+                # Use data URI to avoid HTTP requests during tests
                 "_make_api_request": lambda *args, **kwargs: {
                     "uid": "test-uid-123",
                     "status": "completed",
-                    "image_url": "https://cdn.bannerbear.com/test-image.jpg",
+                    "image_url": "data:image/jpeg;base64,/9j/4AAQSkZJRgABAQAAAQABAAD/2wBDAAgGBgcGBQgHBwcJCQgKDBQNDAsLDBkSEw8UHRofHh0aHBwgJC4nICIsIxwcKDcpLDAxNDQ0Hyc5PTgyPC4zNDL/wAALCAABAAEBAREA/8QAHwAAAQUBAQEBAQEAAAAAAAAAAAECAwQFBgcICQoL/8QAtRAAAgEDAwIEAwUFBAQAAAF9AQIDAAQRBRIhMUEGE1FhByJxFDKBkaEII0KxwRVS0fAkM2JyggkKFhcYGRolJicoKSo0NTY3ODk6Q0RFRkdISUpTVFVWV1hZWmNkZWZnaGlqc3R1dnd4eXqDhIWGh4iJipKTlJWWl5iZmqKjpKWmp6ipqrKztLW2t7i5usLDxMXGx8jJytLT1NXW19jZ2uHi4+Tl5ufo6erx8vP09fb3+Pn6/9oACAEBAAA/APn+v//Z",
                 }
             },
             test_credentials=TEST_CREDENTIALS,
@@ -177,7 +182,12 @@ class BannerbearTextOverlayBlock(Block):
             raise Exception(error_msg)
 
     async def run(
-        self, input_data: Input, *, credentials: APIKeyCredentials, **kwargs
+        self,
+        input_data: Input,
+        *,
+        credentials: APIKeyCredentials,
+        execution_context: ExecutionContext,
+        **kwargs,
     ) -> BlockOutput:
         # Build the modifications array
         modifications = []
@@ -234,6 +244,18 @@ class BannerbearTextOverlayBlock(Block):
 
         # Synchronous request - image should be ready
         yield "success", True
-        yield "image_url", data.get("image_url", "")
+
+        # Store the generated image to workspace for persistence
+        image_url = data.get("image_url", "")
+        if image_url:
+            stored_url = await store_media_file(
+                file=MediaFileType(image_url),
+                execution_context=execution_context,
+                return_format="for_block_output",
+            )
+            yield "image_url", stored_url
+        else:
+            yield "image_url", ""
+
         yield "uid", data.get("uid", "")
         yield "status", data.get("status", "completed")

autogpt_platform/backend/backend/blocks/basic.py

@@ -9,6 +9,7 @@ from backend.data.block import (
     BlockSchemaOutput,
     BlockType,
 )
+from backend.data.execution import ExecutionContext
 from backend.data.model import SchemaField
 from backend.util.file import store_media_file
 from backend.util.type import MediaFileType, convert
@@ -17,10 +18,10 @@ from backend.util.type import MediaFileType, convert
 class FileStoreBlock(Block):
     class Input(BlockSchemaInput):
         file_in: MediaFileType = SchemaField(
-            description="The file to store in the temporary directory, it can be a URL, data URI, or local path."
+            description="The file to download and store. Can be a URL (https://...), data URI, or local path."
         )
         base_64: bool = SchemaField(
-            description="Whether produce an output in base64 format (not recommended, you can pass the string path just fine accross blocks).",
+            description="Whether to produce output in base64 format (not recommended, you can pass the file reference across blocks).",
             default=False,
             advanced=True,
             title="Produce Base64 Output",
@@ -28,13 +29,18 @@ class FileStoreBlock(Block):
 
     class Output(BlockSchemaOutput):
         file_out: MediaFileType = SchemaField(
-            description="The relative path to the stored file in the temporary directory."
+            description="Reference to the stored file. In CoPilot: workspace:// URI (visible in list_workspace_files). In graphs: data URI for passing to other blocks."
         )
 
     def __init__(self):
         super().__init__(
             id="cbb50872-625b-42f0-8203-a2ae78242d8a",
-            description="Stores the input file in the temporary directory.",
+            description=(
+                "Downloads and stores a file from a URL, data URI, or local path. "
+                "Use this to fetch images, documents, or other files for processing. "
+                "In CoPilot: saves to workspace (use list_workspace_files to see it). "
+                "In graphs: outputs a data URI to pass to other blocks."
+            ),
             categories={BlockCategory.BASIC, BlockCategory.MULTIMEDIA},
             input_schema=FileStoreBlock.Input,
             output_schema=FileStoreBlock.Output,
@@ -45,15 +51,18 @@ class FileStoreBlock(Block):
         self,
         input_data: Input,
         *,
-        graph_exec_id: str,
-        user_id: str,
+        execution_context: ExecutionContext,
         **kwargs,
     ) -> BlockOutput:
+        # Determine return format based on user preference
+        # for_external_api: always returns data URI (base64) - honors "Produce Base64 Output"
+        # for_block_output: smart format - workspace:// in CoPilot, data URI in graphs
+        return_format = "for_external_api" if input_data.base_64 else "for_block_output"
+
         yield "file_out", await store_media_file(
-            graph_exec_id=graph_exec_id,
             file=input_data.file_in,
-            user_id=user_id,
-            return_content=input_data.base_64,
+            execution_context=execution_context,
+            return_format=return_format,
         )
 
 
@@ -116,6 +125,7 @@ class PrintToConsoleBlock(Block):
             input_schema=PrintToConsoleBlock.Input,
             output_schema=PrintToConsoleBlock.Output,
             test_input={"text": "Hello, World!"},
+            is_sensitive_action=True,
             test_output=[
                 ("output", "Hello, World!"),
                 ("status", "printed"),

autogpt_platform/backend/backend/blocks/claude_code.py

@@ -0,0 +1,659 @@
+import json
+import shlex
+import uuid
+from typing import Literal, Optional
+
+from e2b import AsyncSandbox as BaseAsyncSandbox
+from pydantic import BaseModel, SecretStr
+
+from backend.data.block import (
+    Block,
+    BlockCategory,
+    BlockOutput,
+    BlockSchemaInput,
+    BlockSchemaOutput,
+)
+from backend.data.model import (
+    APIKeyCredentials,
+    CredentialsField,
+    CredentialsMetaInput,
+    SchemaField,
+)
+from backend.integrations.providers import ProviderName
+
+
+class ClaudeCodeExecutionError(Exception):
+    """Exception raised when Claude Code execution fails.
+
+    Carries the sandbox_id so it can be returned to the user for cleanup
+    when dispose_sandbox=False.
+    """
+
+    def __init__(self, message: str, sandbox_id: str = ""):
+        super().__init__(message)
+        self.sandbox_id = sandbox_id
+
+
+# Test credentials for E2B
+TEST_E2B_CREDENTIALS = APIKeyCredentials(
+    id="01234567-89ab-cdef-0123-456789abcdef",
+    provider="e2b",
+    api_key=SecretStr("mock-e2b-api-key"),
+    title="Mock E2B API key",
+    expires_at=None,
+)
+TEST_E2B_CREDENTIALS_INPUT = {
+    "provider": TEST_E2B_CREDENTIALS.provider,
+    "id": TEST_E2B_CREDENTIALS.id,
+    "type": TEST_E2B_CREDENTIALS.type,
+    "title": TEST_E2B_CREDENTIALS.title,
+}
+
+# Test credentials for Anthropic
+TEST_ANTHROPIC_CREDENTIALS = APIKeyCredentials(
+    id="2e568a2b-b2ea-475a-8564-9a676bf31c56",
+    provider="anthropic",
+    api_key=SecretStr("mock-anthropic-api-key"),
+    title="Mock Anthropic API key",
+    expires_at=None,
+)
+TEST_ANTHROPIC_CREDENTIALS_INPUT = {
+    "provider": TEST_ANTHROPIC_CREDENTIALS.provider,
+    "id": TEST_ANTHROPIC_CREDENTIALS.id,
+    "type": TEST_ANTHROPIC_CREDENTIALS.type,
+    "title": TEST_ANTHROPIC_CREDENTIALS.title,
+}
+
+
+class ClaudeCodeBlock(Block):
+    """
+    Execute tasks using Claude Code (Anthropic's AI coding assistant) in an E2B sandbox.
+
+    Claude Code can create files, install tools, run commands, and perform complex
+    coding tasks autonomously within a secure sandbox environment.
+    """
+
+    # Use base template - we'll install Claude Code ourselves for latest version
+    DEFAULT_TEMPLATE = "base"
+
+    class Input(BlockSchemaInput):
+        e2b_credentials: CredentialsMetaInput[
+            Literal[ProviderName.E2B], Literal["api_key"]
+        ] = CredentialsField(
+            description=(
+                "API key for the E2B platform to create the sandbox. "
+                "Get one on the [e2b website](https://e2b.dev/docs)"
+            ),
+        )
+
+        anthropic_credentials: CredentialsMetaInput[
+            Literal[ProviderName.ANTHROPIC], Literal["api_key"]
+        ] = CredentialsField(
+            description=(
+                "API key for Anthropic to power Claude Code. "
+                "Get one at [Anthropic's website](https://console.anthropic.com)"
+            ),
+        )
+
+        prompt: str = SchemaField(
+            description=(
+                "The task or instruction for Claude Code to execute. "
+                "Claude Code can create files, install packages, run commands, "
+                "and perform complex coding tasks."
+            ),
+            placeholder="Create a hello world index.html file",
+            default="",
+            advanced=False,
+        )
+
+        timeout: int = SchemaField(
+            description=(
+                "Sandbox timeout in seconds. Claude Code tasks can take "
+                "a while, so set this appropriately for your task complexity. "
+                "Note: This only applies when creating a new sandbox. "
+                "When reconnecting to an existing sandbox via sandbox_id, "
+                "the original timeout is retained."
+            ),
+            default=300,  # 5 minutes default
+            advanced=True,
+        )
+
+        setup_commands: list[str] = SchemaField(
+            description=(
+                "Optional shell commands to run before executing Claude Code. "
+                "Useful for installing dependencies or setting up the environment."
+            ),
+            default_factory=list,
+            advanced=True,
+        )
+
+        working_directory: str = SchemaField(
+            description="Working directory for Claude Code to operate in.",
+            default="/home/user",
+            advanced=True,
+        )
+
+        # Session/continuation support
+        session_id: str = SchemaField(
+            description=(
+                "Session ID to resume a previous conversation. "
+                "Leave empty for a new conversation. "
+                "Use the session_id from a previous run to continue that conversation."
+            ),
+            default="",
+            advanced=True,
+        )
+
+        sandbox_id: str = SchemaField(
+            description=(
+                "Sandbox ID to reconnect to an existing sandbox. "
+                "Required when resuming a session (along with session_id). "
+                "Use the sandbox_id from a previous run where dispose_sandbox was False."
+            ),
+            default="",
+            advanced=True,
+        )
+
+        conversation_history: str = SchemaField(
+            description=(
+                "Previous conversation history to continue from. "
+                "Use this to restore context on a fresh sandbox if the previous one timed out. "
+                "Pass the conversation_history output from a previous run."
+            ),
+            default="",
+            advanced=True,
+        )
+
+        dispose_sandbox: bool = SchemaField(
+            description=(
+                "Whether to dispose of the sandbox immediately after execution. "
+                "Set to False if you want to continue the conversation later "
+                "(you'll need both sandbox_id and session_id from the output)."
+            ),
+            default=True,
+            advanced=True,
+        )
+
+    class FileOutput(BaseModel):
+        """A file extracted from the sandbox."""
+
+        path: str
+        relative_path: str  # Path relative to working directory (for GitHub, etc.)
+        name: str
+        content: str
+
+    class Output(BlockSchemaOutput):
+        response: str = SchemaField(
+            description="The output/response from Claude Code execution"
+        )
+        files: list["ClaudeCodeBlock.FileOutput"] = SchemaField(
+            description=(
+                "List of text files created/modified by Claude Code during this execution. "
+                "Each file has 'path', 'relative_path', 'name', and 'content' fields."
+            )
+        )
+        conversation_history: str = SchemaField(
+            description=(
+                "Full conversation history including this turn. "
+                "Pass this to conversation_history input to continue on a fresh sandbox "
+                "if the previous sandbox timed out."
+            )
+        )
+        session_id: str = SchemaField(
+            description=(
+                "Session ID for this conversation. "
+                "Pass this back along with sandbox_id to continue the conversation."
+            )
+        )
+        sandbox_id: Optional[str] = SchemaField(
+            description=(
+                "ID of the sandbox instance. "
+                "Pass this back along with session_id to continue the conversation. "
+                "This is None if dispose_sandbox was True (sandbox was disposed)."
+            ),
+            default=None,
+        )
+        error: str = SchemaField(description="Error message if execution failed")
+
+    def __init__(self):
+        super().__init__(
+            id="4e34f4a5-9b89-4326-ba77-2dd6750b7194",
+            description=(
+                "Execute tasks using Claude Code in an E2B sandbox. "
+                "Claude Code can create files, install tools, run commands, "
+                "and perform complex coding tasks autonomously."
+            ),
+            categories={BlockCategory.DEVELOPER_TOOLS, BlockCategory.AI},
+            input_schema=ClaudeCodeBlock.Input,
+            output_schema=ClaudeCodeBlock.Output,
+            test_credentials={
+                "e2b_credentials": TEST_E2B_CREDENTIALS,
+                "anthropic_credentials": TEST_ANTHROPIC_CREDENTIALS,
+            },
+            test_input={
+                "e2b_credentials": TEST_E2B_CREDENTIALS_INPUT,
+                "anthropic_credentials": TEST_ANTHROPIC_CREDENTIALS_INPUT,
+                "prompt": "Create a hello world HTML file",
+                "timeout": 300,
+                "setup_commands": [],
+                "working_directory": "/home/user",
+                "session_id": "",
+                "sandbox_id": "",
+                "conversation_history": "",
+                "dispose_sandbox": True,
+            },
+            test_output=[
+                ("response", "Created index.html with hello world content"),
+                (
+                    "files",
+                    [
+                        {
+                            "path": "/home/user/index.html",
+                            "relative_path": "index.html",
+                            "name": "index.html",
+                            "content": "<html>Hello World</html>",
+                        }
+                    ],
+                ),
+                (
+                    "conversation_history",
+                    "User: Create a hello world HTML file\n"
+                    "Claude: Created index.html with hello world content",
+                ),
+                ("session_id", str),
+                ("sandbox_id", None),  # None because dispose_sandbox=True in test_input
+            ],
+            test_mock={
+                "execute_claude_code": lambda *args, **kwargs: (
+                    "Created index.html with hello world content",  # response
+                    [
+                        ClaudeCodeBlock.FileOutput(
+                            path="/home/user/index.html",
+                            relative_path="index.html",
+                            name="index.html",
+                            content="<html>Hello World</html>",
+                        )
+                    ],  # files
+                    "User: Create a hello world HTML file\n"
+                    "Claude: Created index.html with hello world content",  # conversation_history
+                    "test-session-id",  # session_id
+                    "sandbox_id",  # sandbox_id
+                ),
+            },
+        )
+
+    async def execute_claude_code(
+        self,
+        e2b_api_key: str,
+        anthropic_api_key: str,
+        prompt: str,
+        timeout: int,
+        setup_commands: list[str],
+        working_directory: str,
+        session_id: str,
+        existing_sandbox_id: str,
+        conversation_history: str,
+        dispose_sandbox: bool,
+    ) -> tuple[str, list["ClaudeCodeBlock.FileOutput"], str, str, str]:
+        """
+        Execute Claude Code in an E2B sandbox.
+
+        Returns:
+            Tuple of (response, files, conversation_history, session_id, sandbox_id)
+        """
+
+        # Validate that sandbox_id is provided when resuming a session
+        if session_id and not existing_sandbox_id:
+            raise ValueError(
+                "sandbox_id is required when resuming a session with session_id. "
+                "The session state is stored in the original sandbox. "
+                "If the sandbox has timed out, use conversation_history instead "
+                "to restore context on a fresh sandbox."
+            )
+
+        sandbox = None
+        sandbox_id = ""
+
+        try:
+            # Either reconnect to existing sandbox or create a new one
+            if existing_sandbox_id:
+                # Reconnect to existing sandbox for conversation continuation
+                sandbox = await BaseAsyncSandbox.connect(
+                    sandbox_id=existing_sandbox_id,
+                    api_key=e2b_api_key,
+                )
+            else:
+                # Create new sandbox
+                sandbox = await BaseAsyncSandbox.create(
+                    template=self.DEFAULT_TEMPLATE,
+                    api_key=e2b_api_key,
+                    timeout=timeout,
+                    envs={"ANTHROPIC_API_KEY": anthropic_api_key},
+                )
+
+                # Install Claude Code from npm (ensures we get the latest version)
+                install_result = await sandbox.commands.run(
+                    "npm install -g @anthropic-ai/claude-code@latest",
+                    timeout=120,  # 2 min timeout for install
+                )
+                if install_result.exit_code != 0:
+                    raise Exception(
+                        f"Failed to install Claude Code: {install_result.stderr}"
+                    )
+
+                # Run any user-provided setup commands
+                for cmd in setup_commands:
+                    setup_result = await sandbox.commands.run(cmd)
+                    if setup_result.exit_code != 0:
+                        raise Exception(
+                            f"Setup command failed: {cmd}\n"
+                            f"Exit code: {setup_result.exit_code}\n"
+                            f"Stdout: {setup_result.stdout}\n"
+                            f"Stderr: {setup_result.stderr}"
+                        )
+
+            # Capture sandbox_id immediately after creation/connection
+            # so it's available for error recovery if dispose_sandbox=False
+            sandbox_id = sandbox.sandbox_id
+
+            # Generate or use provided session ID
+            current_session_id = session_id if session_id else str(uuid.uuid4())
+
+            # Build base Claude flags
+            base_flags = "-p --dangerously-skip-permissions --output-format json"
+
+            # Add conversation history context if provided (for fresh sandbox continuation)
+            history_flag = ""
+            if conversation_history and not session_id:
+                # Inject previous conversation as context via system prompt
+                # Use consistent escaping via _escape_prompt helper
+                escaped_history = self._escape_prompt(
+                    f"Previous conversation context: {conversation_history}"
+                )
+                history_flag = f" --append-system-prompt {escaped_history}"
+
+            # Build Claude command based on whether we're resuming or starting new
+            # Use shlex.quote for working_directory and session IDs to prevent injection
+            safe_working_dir = shlex.quote(working_directory)
+            if session_id:
+                # Resuming existing session (sandbox still alive)
+                safe_session_id = shlex.quote(session_id)
+                claude_command = (
+                    f"cd {safe_working_dir} && "
+                    f"echo {self._escape_prompt(prompt)} | "
+                    f"claude --resume {safe_session_id} {base_flags}"
+                )
+            else:
+                # New session with specific ID
+                safe_current_session_id = shlex.quote(current_session_id)
+                claude_command = (
+                    f"cd {safe_working_dir} && "
+                    f"echo {self._escape_prompt(prompt)} | "
+                    f"claude --session-id {safe_current_session_id} {base_flags}{history_flag}"
+                )
+
+            # Capture timestamp before running Claude Code to filter files later
+            # Capture timestamp 1 second in the past to avoid race condition with file creation
+            timestamp_result = await sandbox.commands.run(
+                "date -u -d '1 second ago' +%Y-%m-%dT%H:%M:%S"
+            )
+            if timestamp_result.exit_code != 0:
+                raise RuntimeError(
+                    f"Failed to capture timestamp: {timestamp_result.stderr}"
+                )
+            start_timestamp = (
+                timestamp_result.stdout.strip() if timestamp_result.stdout else None
+            )
+
+            result = await sandbox.commands.run(
+                claude_command,
+                timeout=0,  # No command timeout - let sandbox timeout handle it
+            )
+
+            # Check for command failure
+            if result.exit_code != 0:
+                error_msg = result.stderr or result.stdout or "Unknown error"
+                raise Exception(
+                    f"Claude Code command failed with exit code {result.exit_code}:\n"
+                    f"{error_msg}"
+                )
+
+            raw_output = result.stdout or ""
+
+            # Parse JSON output to extract response and build conversation history
+            response = ""
+            new_conversation_history = conversation_history or ""
+
+            try:
+                # The JSON output contains the result
+                output_data = json.loads(raw_output)
+                response = output_data.get("result", raw_output)
+
+                # Build conversation history entry
+                turn_entry = f"User: {prompt}\nClaude: {response}"
+                if new_conversation_history:
+                    new_conversation_history = (
+                        f"{new_conversation_history}\n\n{turn_entry}"
+                    )
+                else:
+                    new_conversation_history = turn_entry
+
+            except json.JSONDecodeError:
+                # If not valid JSON, use raw output
+                response = raw_output
+                turn_entry = f"User: {prompt}\nClaude: {response}"
+                if new_conversation_history:
+                    new_conversation_history = (
+                        f"{new_conversation_history}\n\n{turn_entry}"
+                    )
+                else:
+                    new_conversation_history = turn_entry
+
+            # Extract files created/modified during this run
+            files = await self._extract_files(
+                sandbox, working_directory, start_timestamp
+            )
+
+            return (
+                response,
+                files,
+                new_conversation_history,
+                current_session_id,
+                sandbox_id,
+            )
+
+        except Exception as e:
+            # Wrap exception with sandbox_id so caller can access/cleanup
+            # the preserved sandbox when dispose_sandbox=False
+            raise ClaudeCodeExecutionError(str(e), sandbox_id) from e
+
+        finally:
+            if dispose_sandbox and sandbox:
+                await sandbox.kill()
+
+    async def _extract_files(
+        self,
+        sandbox: BaseAsyncSandbox,
+        working_directory: str,
+        since_timestamp: str | None = None,
+    ) -> list["ClaudeCodeBlock.FileOutput"]:
+        """
+        Extract text files created/modified during this Claude Code execution.
+
+        Args:
+            sandbox: The E2B sandbox instance
+            working_directory: Directory to search for files
+            since_timestamp: ISO timestamp - only return files modified after this time
+
+        Returns:
+            List of FileOutput objects with path, relative_path, name, and content
+        """
+        files: list[ClaudeCodeBlock.FileOutput] = []
+
+        # Text file extensions we can safely read as text
+        text_extensions = {
+            ".txt",
+            ".md",
+            ".html",
+            ".htm",
+            ".css",
+            ".js",
+            ".ts",
+            ".jsx",
+            ".tsx",
+            ".json",
+            ".xml",
+            ".yaml",
+            ".yml",
+            ".toml",
+            ".ini",
+            ".cfg",
+            ".conf",
+            ".py",
+            ".rb",
+            ".php",
+            ".java",
+            ".c",
+            ".cpp",
+            ".h",
+            ".hpp",
+            ".cs",
+            ".go",
+            ".rs",
+            ".swift",
+            ".kt",
+            ".scala",
+            ".sh",
+            ".bash",
+            ".zsh",
+            ".sql",
+            ".graphql",
+            ".env",
+            ".gitignore",
+            ".dockerfile",
+            "Dockerfile",
+            ".vue",
+            ".svelte",
+            ".astro",
+            ".mdx",
+            ".rst",
+            ".tex",
+            ".csv",
+            ".log",
+        }
+
+        try:
+            # List files recursively using find command
+            # Exclude node_modules and .git directories, but allow hidden files
+            # like .env and .gitignore (they're filtered by text_extensions later)
+            # Filter by timestamp to only get files created/modified during this run
+            safe_working_dir = shlex.quote(working_directory)
+            timestamp_filter = ""
+            if since_timestamp:
+                timestamp_filter = f"-newermt {shlex.quote(since_timestamp)} "
+            find_result = await sandbox.commands.run(
+                f"find {safe_working_dir} -type f "
+                f"{timestamp_filter}"
+                f"-not -path '*/node_modules/*' "
+                f"-not -path '*/.git/*' "
+                f"2>/dev/null"
+            )
+
+            if find_result.stdout:
+                for file_path in find_result.stdout.strip().split("\n"):
+                    if not file_path:
+                        continue
+
+                    # Check if it's a text file we can read
+                    is_text = any(
+                        file_path.endswith(ext) for ext in text_extensions
+                    ) or file_path.endswith("Dockerfile")
+
+                    if is_text:
+                        try:
+                            content = await sandbox.files.read(file_path)
+                            # Handle bytes or string
+                            if isinstance(content, bytes):
+                                content = content.decode("utf-8", errors="replace")
+
+                            # Extract filename from path
+                            file_name = file_path.split("/")[-1]
+
+                            # Calculate relative path by stripping working directory
+                            relative_path = file_path
+                            if file_path.startswith(working_directory):
+                                relative_path = file_path[len(working_directory) :]
+                                # Remove leading slash if present
+                                if relative_path.startswith("/"):
+                                    relative_path = relative_path[1:]
+
+                            files.append(
+                                ClaudeCodeBlock.FileOutput(
+                                    path=file_path,
+                                    relative_path=relative_path,
+                                    name=file_name,
+                                    content=content,
+                                )
+                            )
+                        except Exception:
+                            # Skip files that can't be read
+                            pass
+
+        except Exception:
+            # If file extraction fails, return empty results
+            pass
+
+        return files
+
+    def _escape_prompt(self, prompt: str) -> str:
+        """Escape the prompt for safe shell execution."""
+        # Use single quotes and escape any single quotes in the prompt
+        escaped = prompt.replace("'", "'\"'\"'")
+        return f"'{escaped}'"
+
+    async def run(
+        self,
+        input_data: Input,
+        *,
+        e2b_credentials: APIKeyCredentials,
+        anthropic_credentials: APIKeyCredentials,
+        **kwargs,
+    ) -> BlockOutput:
+        try:
+            (
+                response,
+                files,
+                conversation_history,
+                session_id,
+                sandbox_id,
+            ) = await self.execute_claude_code(
+                e2b_api_key=e2b_credentials.api_key.get_secret_value(),
+                anthropic_api_key=anthropic_credentials.api_key.get_secret_value(),
+                prompt=input_data.prompt,
+                timeout=input_data.timeout,
+                setup_commands=input_data.setup_commands,
+                working_directory=input_data.working_directory,
+                session_id=input_data.session_id,
+                existing_sandbox_id=input_data.sandbox_id,
+                conversation_history=input_data.conversation_history,
+                dispose_sandbox=input_data.dispose_sandbox,
+            )
+
+            yield "response", response
+            # Always yield files (empty list if none) to match Output schema
+            yield "files", [f.model_dump() for f in files]
+            # Always yield conversation_history so user can restore context on fresh sandbox
+            yield "conversation_history", conversation_history
+            # Always yield session_id so user can continue conversation
+            yield "session_id", session_id
+            # Always yield sandbox_id (None if disposed) to match Output schema
+            yield "sandbox_id", sandbox_id if not input_data.dispose_sandbox else None
+
+        except ClaudeCodeExecutionError as e:
+            yield "error", str(e)
+            # If sandbox was preserved (dispose_sandbox=False), yield sandbox_id
+            # so user can reconnect to or clean up the orphaned sandbox
+            if not input_data.dispose_sandbox and e.sandbox_id:
+                yield "sandbox_id", e.sandbox_id
+        except Exception as e:
+            yield "error", str(e)

autogpt_platform/backend/backend/blocks/discord/bot_blocks.py

@@ -15,6 +15,7 @@ from backend.data.block import (
     BlockSchemaInput,
     BlockSchemaOutput,
 )
+from backend.data.execution import ExecutionContext
 from backend.data.model import APIKeyCredentials, SchemaField
 from backend.util.file import store_media_file
 from backend.util.request import Requests
@@ -666,8 +667,7 @@ class SendDiscordFileBlock(Block):
         file: MediaFileType,
         filename: str,
         message_content: str,
-        graph_exec_id: str,
-        user_id: str,
+        execution_context: ExecutionContext,
     ) -> dict:
         intents = discord.Intents.default()
         intents.guilds = True
@@ -731,10 +731,9 @@ class SendDiscordFileBlock(Block):
                     # Local file path - read from stored media file
                     # This would be a path from a previous block's output
                     stored_file = await store_media_file(
-                        graph_exec_id=graph_exec_id,
                         file=file,
-                        user_id=user_id,
-                        return_content=True,  # Get as data URI
+                        execution_context=execution_context,
+                        return_format="for_external_api",  # Get content to send to Discord
                     )
                     # Now process as data URI
                     header, encoded = stored_file.split(",", 1)
@@ -781,8 +780,7 @@ class SendDiscordFileBlock(Block):
         input_data: Input,
         *,
         credentials: APIKeyCredentials,
-        graph_exec_id: str,
-        user_id: str,
+        execution_context: ExecutionContext,
         **kwargs,
     ) -> BlockOutput:
         try:
@@ -793,8 +791,7 @@ class SendDiscordFileBlock(Block):
                 file=input_data.file,
                 filename=input_data.filename,
                 message_content=input_data.message_content,
-                graph_exec_id=graph_exec_id,
-                user_id=user_id,
+                execution_context=execution_context,
             )
 
             yield "status", result.get("status", "Unknown error")

autogpt_platform/backend/backend/blocks/elevenlabs/_auth.py

@@ -0,0 +1,28 @@
+"""ElevenLabs integration blocks - test credentials and shared utilities."""
+
+from typing import Literal
+
+from pydantic import SecretStr
+
+from backend.data.model import APIKeyCredentials, CredentialsMetaInput
+from backend.integrations.providers import ProviderName
+
+TEST_CREDENTIALS = APIKeyCredentials(
+    id="01234567-89ab-cdef-0123-456789abcdef",
+    provider="elevenlabs",
+    api_key=SecretStr("mock-elevenlabs-api-key"),
+    title="Mock ElevenLabs API key",
+    expires_at=None,
+)
+
+TEST_CREDENTIALS_INPUT = {
+    "provider": TEST_CREDENTIALS.provider,
+    "id": TEST_CREDENTIALS.id,
+    "type": TEST_CREDENTIALS.type,
+    "title": TEST_CREDENTIALS.title,
+}
+
+ElevenLabsCredentials = APIKeyCredentials
+ElevenLabsCredentialsInput = CredentialsMetaInput[
+    Literal[ProviderName.ELEVENLABS], Literal["api_key"]
+]

autogpt_platform/backend/backend/blocks/encoder_block.py

@@ -0,0 +1,77 @@
+"""Text encoding block for converting special characters to escape sequences."""
+
+import codecs
+
+from backend.data.block import (
+    Block,
+    BlockCategory,
+    BlockOutput,
+    BlockSchemaInput,
+    BlockSchemaOutput,
+)
+from backend.data.model import SchemaField
+
+
+class TextEncoderBlock(Block):
+    """
+    Encodes a string by converting special characters into escape sequences.
+
+    This block is the inverse of TextDecoderBlock. It takes text containing
+    special characters (like newlines, tabs, etc.) and converts them into
+    their escape sequence representations (e.g., newline becomes \\n).
+    """
+
+    class Input(BlockSchemaInput):
+        """Input schema for TextEncoderBlock."""
+
+        text: str = SchemaField(
+            description="A string containing special characters to be encoded",
+            placeholder="Your text with newlines and quotes to encode",
+        )
+
+    class Output(BlockSchemaOutput):
+        """Output schema for TextEncoderBlock."""
+
+        encoded_text: str = SchemaField(
+            description="The encoded text with special characters converted to escape sequences"
+        )
+        error: str = SchemaField(description="Error message if encoding fails")
+
+    def __init__(self):
+        super().__init__(
+            id="5185f32e-4b65-4ecf-8fbb-873f003f09d6",
+            description="Encodes a string by converting special characters into escape sequences",
+            categories={BlockCategory.TEXT},
+            input_schema=TextEncoderBlock.Input,
+            output_schema=TextEncoderBlock.Output,
+            test_input={
+                "text": """Hello
+World!
+This is a "quoted" string."""
+            },
+            test_output=[
+                (
+                    "encoded_text",
+                    """Hello\\nWorld!\\nThis is a "quoted" string.""",
+                )
+            ],
+        )
+
+    async def run(self, input_data: Input, **kwargs) -> BlockOutput:
+        """
+        Encode the input text by converting special characters to escape sequences.
+
+        Args:
+            input_data: The input containing the text to encode.
+            **kwargs: Additional keyword arguments (unused).
+
+        Yields:
+            The encoded text with escape sequences, or an error message if encoding fails.
+        """
+        try:
+            encoded_text = codecs.encode(input_data.text, "unicode_escape").decode(
+                "utf-8"
+            )
+            yield "encoded_text", encoded_text
+        except Exception as e:
+            yield "error", f"Encoding error: {str(e)}"

autogpt_platform/backend/backend/blocks/exa/websets.py

@@ -478,7 +478,7 @@ class ExaCreateOrFindWebsetBlock(Block):
         aexa = AsyncExa(api_key=credentials.api_key.get_secret_value())
 
         try:
-            webset = aexa.websets.get(id=input_data.external_id)
+            webset = await aexa.websets.get(id=input_data.external_id)
             webset_result = Webset.model_validate(webset.model_dump(by_alias=True))
 
             yield "webset", webset_result
@@ -494,7 +494,7 @@ class ExaCreateOrFindWebsetBlock(Block):
                         count=input_data.search_count,
                     )
 
-                webset = aexa.websets.create(
+                webset = await aexa.websets.create(
                     params=CreateWebsetParameters(
                         search=search_params,
                         external_id=input_data.external_id,
@@ -554,7 +554,7 @@ class ExaUpdateWebsetBlock(Block):
         if input_data.metadata is not None:
             payload["metadata"] = input_data.metadata
 
-        sdk_webset = aexa.websets.update(id=input_data.webset_id, params=payload)
+        sdk_webset = await aexa.websets.update(id=input_data.webset_id, params=payload)
 
         status_str = (
             sdk_webset.status.value
@@ -617,7 +617,7 @@ class ExaListWebsetsBlock(Block):
     ) -> BlockOutput:
         aexa = AsyncExa(api_key=credentials.api_key.get_secret_value())
 
-        response = aexa.websets.list(
+        response = await aexa.websets.list(
             cursor=input_data.cursor,
             limit=input_data.limit,
         )
@@ -678,7 +678,7 @@ class ExaGetWebsetBlock(Block):
     ) -> BlockOutput:
         aexa = AsyncExa(api_key=credentials.api_key.get_secret_value())
 
-        sdk_webset = aexa.websets.get(id=input_data.webset_id)
+        sdk_webset = await aexa.websets.get(id=input_data.webset_id)
 
         status_str = (
             sdk_webset.status.value
@@ -748,7 +748,7 @@ class ExaDeleteWebsetBlock(Block):
     ) -> BlockOutput:
         aexa = AsyncExa(api_key=credentials.api_key.get_secret_value())
 
-        deleted_webset = aexa.websets.delete(id=input_data.webset_id)
+        deleted_webset = await aexa.websets.delete(id=input_data.webset_id)
 
         status_str = (
             deleted_webset.status.value
@@ -798,7 +798,7 @@ class ExaCancelWebsetBlock(Block):
     ) -> BlockOutput:
         aexa = AsyncExa(api_key=credentials.api_key.get_secret_value())
 
-        canceled_webset = aexa.websets.cancel(id=input_data.webset_id)
+        canceled_webset = await aexa.websets.cancel(id=input_data.webset_id)
 
         status_str = (
             canceled_webset.status.value
@@ -968,7 +968,7 @@ class ExaPreviewWebsetBlock(Block):
                 entity["description"] = input_data.entity_description
             payload["entity"] = entity
 
-        sdk_preview = aexa.websets.preview(params=payload)
+        sdk_preview = await aexa.websets.preview(params=payload)
 
         preview = PreviewWebsetModel.from_sdk(sdk_preview)
 
@@ -1051,7 +1051,7 @@ class ExaWebsetStatusBlock(Block):
     ) -> BlockOutput:
         aexa = AsyncExa(api_key=credentials.api_key.get_secret_value())
 
-        webset = aexa.websets.get(id=input_data.webset_id)
+        webset = await aexa.websets.get(id=input_data.webset_id)
 
         status = (
             webset.status.value
@@ -1185,7 +1185,7 @@ class ExaWebsetSummaryBlock(Block):
     ) -> BlockOutput:
         aexa = AsyncExa(api_key=credentials.api_key.get_secret_value())
 
-        webset = aexa.websets.get(id=input_data.webset_id)
+        webset = await aexa.websets.get(id=input_data.webset_id)
 
         # Extract basic info
         webset_id = webset.id
@@ -1211,7 +1211,7 @@ class ExaWebsetSummaryBlock(Block):
         total_items = 0
 
         if input_data.include_sample_items and input_data.sample_size > 0:
-            items_response = aexa.websets.items.list(
+            items_response = await aexa.websets.items.list(
                 webset_id=input_data.webset_id, limit=input_data.sample_size
             )
             sample_items_data = [
@@ -1362,7 +1362,7 @@ class ExaWebsetReadyCheckBlock(Block):
         aexa = AsyncExa(api_key=credentials.api_key.get_secret_value())
 
         # Get webset details
-        webset = aexa.websets.get(id=input_data.webset_id)
+        webset = await aexa.websets.get(id=input_data.webset_id)
 
         status = (
             webset.status.value

autogpt_platform/backend/backend/blocks/exa/websets_enrichment.py

@@ -202,7 +202,7 @@ class ExaCreateEnrichmentBlock(Block):
         # Use AsyncExa SDK
         aexa = AsyncExa(api_key=credentials.api_key.get_secret_value())
 
-        sdk_enrichment = aexa.websets.enrichments.create(
+        sdk_enrichment = await aexa.websets.enrichments.create(
             webset_id=input_data.webset_id, params=payload
         )
 
@@ -223,7 +223,7 @@ class ExaCreateEnrichmentBlock(Block):
             items_enriched = 0
 
             while time.time() - poll_start < input_data.polling_timeout:
-                current_enrich = aexa.websets.enrichments.get(
+                current_enrich = await aexa.websets.enrichments.get(
                     webset_id=input_data.webset_id, id=enrichment_id
                 )
                 current_status = (
@@ -234,7 +234,7 @@ class ExaCreateEnrichmentBlock(Block):
 
                 if current_status in ["completed", "failed", "cancelled"]:
                     # Estimate items from webset searches
-                    webset = aexa.websets.get(id=input_data.webset_id)
+                    webset = await aexa.websets.get(id=input_data.webset_id)
                     if webset.searches:
                         for search in webset.searches:
                             if search.progress:
@@ -329,7 +329,7 @@ class ExaGetEnrichmentBlock(Block):
         # Use AsyncExa SDK
         aexa = AsyncExa(api_key=credentials.api_key.get_secret_value())
 
-        sdk_enrichment = aexa.websets.enrichments.get(
+        sdk_enrichment = await aexa.websets.enrichments.get(
             webset_id=input_data.webset_id, id=input_data.enrichment_id
         )
 
@@ -474,7 +474,7 @@ class ExaDeleteEnrichmentBlock(Block):
         # Use AsyncExa SDK
         aexa = AsyncExa(api_key=credentials.api_key.get_secret_value())
 
-        deleted_enrichment = aexa.websets.enrichments.delete(
+        deleted_enrichment = await aexa.websets.enrichments.delete(
             webset_id=input_data.webset_id, id=input_data.enrichment_id
         )
 
@@ -525,13 +525,13 @@ class ExaCancelEnrichmentBlock(Block):
         # Use AsyncExa SDK
         aexa = AsyncExa(api_key=credentials.api_key.get_secret_value())
 
-        canceled_enrichment = aexa.websets.enrichments.cancel(
+        canceled_enrichment = await aexa.websets.enrichments.cancel(
             webset_id=input_data.webset_id, id=input_data.enrichment_id
         )
 
         # Try to estimate how many items were enriched before cancellation
         items_enriched = 0
-        items_response = aexa.websets.items.list(
+        items_response = await aexa.websets.items.list(
             webset_id=input_data.webset_id, limit=100
         )
 

autogpt_platform/backend/backend/blocks/exa/websets_import_export.py

@@ -222,7 +222,7 @@ class ExaCreateImportBlock(Block):
     def _create_test_mock():
         """Create test mocks for the AsyncExa SDK."""
         from datetime import datetime
-        from unittest.mock import MagicMock
+        from unittest.mock import AsyncMock, MagicMock
 
         # Create mock SDK import object
         mock_import = MagicMock()
@@ -247,7 +247,7 @@ class ExaCreateImportBlock(Block):
         return {
             "_get_client": lambda *args, **kwargs: MagicMock(
                 websets=MagicMock(
-                    imports=MagicMock(create=lambda *args, **kwargs: mock_import)
+                    imports=MagicMock(create=AsyncMock(return_value=mock_import))
                 )
             )
         }
@@ -294,7 +294,7 @@ class ExaCreateImportBlock(Block):
         if input_data.metadata:
             payload["metadata"] = input_data.metadata
 
-        sdk_import = aexa.websets.imports.create(
+        sdk_import = await aexa.websets.imports.create(
             params=payload, csv_data=input_data.csv_data
         )
 
@@ -360,7 +360,7 @@ class ExaGetImportBlock(Block):
         # Use AsyncExa SDK
         aexa = AsyncExa(api_key=credentials.api_key.get_secret_value())
 
-        sdk_import = aexa.websets.imports.get(import_id=input_data.import_id)
+        sdk_import = await aexa.websets.imports.get(import_id=input_data.import_id)
 
         import_obj = ImportModel.from_sdk(sdk_import)
 
@@ -426,7 +426,7 @@ class ExaListImportsBlock(Block):
         # Use AsyncExa SDK
         aexa = AsyncExa(api_key=credentials.api_key.get_secret_value())
 
-        response = aexa.websets.imports.list(
+        response = await aexa.websets.imports.list(
             cursor=input_data.cursor,
             limit=input_data.limit,
         )
@@ -474,7 +474,9 @@ class ExaDeleteImportBlock(Block):
         # Use AsyncExa SDK
         aexa = AsyncExa(api_key=credentials.api_key.get_secret_value())
 
-        deleted_import = aexa.websets.imports.delete(import_id=input_data.import_id)
+        deleted_import = await aexa.websets.imports.delete(
+            import_id=input_data.import_id
+        )
 
         yield "import_id", deleted_import.id
         yield "success", "true"
@@ -573,14 +575,14 @@ class ExaExportWebsetBlock(Block):
             }
         )
 
-        # Create mock iterator
-        mock_items = [mock_item1, mock_item2]
+        # Create async iterator for list_all
+        async def async_item_iterator(*args, **kwargs):
+            for item in [mock_item1, mock_item2]:
+                yield item
 
         return {
             "_get_client": lambda *args, **kwargs: MagicMock(
-                websets=MagicMock(
-                    items=MagicMock(list_all=lambda *args, **kwargs: iter(mock_items))
-                )
+                websets=MagicMock(items=MagicMock(list_all=async_item_iterator))
             )
         }
 
@@ -602,7 +604,7 @@ class ExaExportWebsetBlock(Block):
                 webset_id=input_data.webset_id, limit=input_data.max_items
             )
 
-            for sdk_item in item_iterator:
+            async for sdk_item in item_iterator:
                 if len(all_items) >= input_data.max_items:
                     break
 

autogpt_platform/backend/backend/blocks/exa/websets_items.py

@@ -178,7 +178,7 @@ class ExaGetWebsetItemBlock(Block):
     ) -> BlockOutput:
         aexa = AsyncExa(api_key=credentials.api_key.get_secret_value())
 
-        sdk_item = aexa.websets.items.get(
+        sdk_item = await aexa.websets.items.get(
             webset_id=input_data.webset_id, id=input_data.item_id
         )
 
@@ -269,7 +269,7 @@ class ExaListWebsetItemsBlock(Block):
             response = None
 
             while time.time() - start_time < input_data.wait_timeout:
-                response = aexa.websets.items.list(
+                response = await aexa.websets.items.list(
                     webset_id=input_data.webset_id,
                     cursor=input_data.cursor,
                     limit=input_data.limit,
@@ -282,13 +282,13 @@ class ExaListWebsetItemsBlock(Block):
                 interval = min(interval * 1.2, 10)
 
             if not response:
-                response = aexa.websets.items.list(
+                response = await aexa.websets.items.list(
                     webset_id=input_data.webset_id,
                     cursor=input_data.cursor,
                     limit=input_data.limit,
                 )
         else:
-            response = aexa.websets.items.list(
+            response = await aexa.websets.items.list(
                 webset_id=input_data.webset_id,
                 cursor=input_data.cursor,
                 limit=input_data.limit,
@@ -340,7 +340,7 @@ class ExaDeleteWebsetItemBlock(Block):
     ) -> BlockOutput:
         aexa = AsyncExa(api_key=credentials.api_key.get_secret_value())
 
-        deleted_item = aexa.websets.items.delete(
+        deleted_item = await aexa.websets.items.delete(
             webset_id=input_data.webset_id, id=input_data.item_id
         )
 
@@ -408,7 +408,7 @@ class ExaBulkWebsetItemsBlock(Block):
             webset_id=input_data.webset_id, limit=input_data.max_items
         )
 
-        for sdk_item in item_iterator:
+        async for sdk_item in item_iterator:
             if len(all_items) >= input_data.max_items:
                 break
 
@@ -475,7 +475,7 @@ class ExaWebsetItemsSummaryBlock(Block):
         # Use AsyncExa SDK
         aexa = AsyncExa(api_key=credentials.api_key.get_secret_value())
 
-        webset = aexa.websets.get(id=input_data.webset_id)
+        webset = await aexa.websets.get(id=input_data.webset_id)
 
         entity_type = "unknown"
         if webset.searches:
@@ -495,7 +495,7 @@ class ExaWebsetItemsSummaryBlock(Block):
         # Get sample items if requested
         sample_items: List[WebsetItemModel] = []
         if input_data.sample_size > 0:
-            items_response = aexa.websets.items.list(
+            items_response = await aexa.websets.items.list(
                 webset_id=input_data.webset_id, limit=input_data.sample_size
             )
             # Convert to our stable models
@@ -569,7 +569,7 @@ class ExaGetNewItemsBlock(Block):
         aexa = AsyncExa(api_key=credentials.api_key.get_secret_value())
 
         # Get items starting from cursor
-        response = aexa.websets.items.list(
+        response = await aexa.websets.items.list(
             webset_id=input_data.webset_id,
             cursor=input_data.since_cursor,
             limit=input_data.max_items,

autogpt_platform/backend/backend/blocks/exa/websets_monitor.py

@@ -233,7 +233,7 @@ class ExaCreateMonitorBlock(Block):
     def _create_test_mock():
         """Create test mocks for the AsyncExa SDK."""
         from datetime import datetime
-        from unittest.mock import MagicMock
+        from unittest.mock import AsyncMock, MagicMock
 
         # Create mock SDK monitor object
         mock_monitor = MagicMock()
@@ -263,7 +263,7 @@ class ExaCreateMonitorBlock(Block):
         return {
             "_get_client": lambda *args, **kwargs: MagicMock(
                 websets=MagicMock(
-                    monitors=MagicMock(create=lambda *args, **kwargs: mock_monitor)
+                    monitors=MagicMock(create=AsyncMock(return_value=mock_monitor))
                 )
             )
         }
@@ -320,7 +320,7 @@ class ExaCreateMonitorBlock(Block):
         if input_data.metadata:
             payload["metadata"] = input_data.metadata
 
-        sdk_monitor = aexa.websets.monitors.create(params=payload)
+        sdk_monitor = await aexa.websets.monitors.create(params=payload)
 
         monitor = MonitorModel.from_sdk(sdk_monitor)
 
@@ -384,7 +384,7 @@ class ExaGetMonitorBlock(Block):
         # Use AsyncExa SDK
         aexa = AsyncExa(api_key=credentials.api_key.get_secret_value())
 
-        sdk_monitor = aexa.websets.monitors.get(monitor_id=input_data.monitor_id)
+        sdk_monitor = await aexa.websets.monitors.get(monitor_id=input_data.monitor_id)
 
         monitor = MonitorModel.from_sdk(sdk_monitor)
 
@@ -476,7 +476,7 @@ class ExaUpdateMonitorBlock(Block):
         if input_data.metadata is not None:
             payload["metadata"] = input_data.metadata
 
-        sdk_monitor = aexa.websets.monitors.update(
+        sdk_monitor = await aexa.websets.monitors.update(
             monitor_id=input_data.monitor_id, params=payload
         )
 
@@ -522,7 +522,9 @@ class ExaDeleteMonitorBlock(Block):
         # Use AsyncExa SDK
         aexa = AsyncExa(api_key=credentials.api_key.get_secret_value())
 
-        deleted_monitor = aexa.websets.monitors.delete(monitor_id=input_data.monitor_id)
+        deleted_monitor = await aexa.websets.monitors.delete(
+            monitor_id=input_data.monitor_id
+        )
 
         yield "monitor_id", deleted_monitor.id
         yield "success", "true"
@@ -579,7 +581,7 @@ class ExaListMonitorsBlock(Block):
         # Use AsyncExa SDK
         aexa = AsyncExa(api_key=credentials.api_key.get_secret_value())
 
-        response = aexa.websets.monitors.list(
+        response = await aexa.websets.monitors.list(
             cursor=input_data.cursor,
             limit=input_data.limit,
             webset_id=input_data.webset_id,

autogpt_platform/backend/backend/blocks/exa/websets_polling.py

@@ -121,7 +121,7 @@ class ExaWaitForWebsetBlock(Block):
                 WebsetTargetStatus.IDLE,
                 WebsetTargetStatus.ANY_COMPLETE,
             ]:
-                final_webset = aexa.websets.wait_until_idle(
+                final_webset = await aexa.websets.wait_until_idle(
                     id=input_data.webset_id,
                     timeout=input_data.timeout,
                     poll_interval=input_data.check_interval,
@@ -164,7 +164,7 @@ class ExaWaitForWebsetBlock(Block):
                 interval = input_data.check_interval
                 while time.time() - start_time < input_data.timeout:
                     # Get current webset status
-                    webset = aexa.websets.get(id=input_data.webset_id)
+                    webset = await aexa.websets.get(id=input_data.webset_id)
                     current_status = (
                         webset.status.value
                         if hasattr(webset.status, "value")
@@ -209,7 +209,7 @@ class ExaWaitForWebsetBlock(Block):
 
                 # Timeout reached
                 elapsed = time.time() - start_time
-                webset = aexa.websets.get(id=input_data.webset_id)
+                webset = await aexa.websets.get(id=input_data.webset_id)
                 final_status = (
                     webset.status.value
                     if hasattr(webset.status, "value")
@@ -345,7 +345,7 @@ class ExaWaitForSearchBlock(Block):
         try:
             while time.time() - start_time < input_data.timeout:
                 # Get current search status using SDK
-                search = aexa.websets.searches.get(
+                search = await aexa.websets.searches.get(
                     webset_id=input_data.webset_id, id=input_data.search_id
                 )
 
@@ -401,7 +401,7 @@ class ExaWaitForSearchBlock(Block):
             elapsed = time.time() - start_time
 
             # Get last known status
-            search = aexa.websets.searches.get(
+            search = await aexa.websets.searches.get(
                 webset_id=input_data.webset_id, id=input_data.search_id
             )
             final_status = (
@@ -503,7 +503,7 @@ class ExaWaitForEnrichmentBlock(Block):
         try:
             while time.time() - start_time < input_data.timeout:
                 # Get current enrichment status using SDK
-                enrichment = aexa.websets.enrichments.get(
+                enrichment = await aexa.websets.enrichments.get(
                     webset_id=input_data.webset_id, id=input_data.enrichment_id
                 )
 
@@ -548,7 +548,7 @@ class ExaWaitForEnrichmentBlock(Block):
             elapsed = time.time() - start_time
 
             # Get last known status
-            enrichment = aexa.websets.enrichments.get(
+            enrichment = await aexa.websets.enrichments.get(
                 webset_id=input_data.webset_id, id=input_data.enrichment_id
             )
             final_status = (
@@ -575,7 +575,7 @@ class ExaWaitForEnrichmentBlock(Block):
     ) -> tuple[list[SampleEnrichmentModel], int]:
         """Get sample enriched data and count."""
         # Get a few items to see enrichment results using SDK
-        response = aexa.websets.items.list(webset_id=webset_id, limit=5)
+        response = await aexa.websets.items.list(webset_id=webset_id, limit=5)
 
         sample_data: list[SampleEnrichmentModel] = []
         enriched_count = 0

autogpt_platform/backend/backend/blocks/exa/websets_search.py

@@ -317,7 +317,7 @@ class ExaCreateWebsetSearchBlock(Block):
 
         aexa = AsyncExa(api_key=credentials.api_key.get_secret_value())
 
-        sdk_search = aexa.websets.searches.create(
+        sdk_search = await aexa.websets.searches.create(
             webset_id=input_data.webset_id, params=payload
         )
 
@@ -350,7 +350,7 @@ class ExaCreateWebsetSearchBlock(Block):
             poll_start = time.time()
 
             while time.time() - poll_start < input_data.polling_timeout:
-                current_search = aexa.websets.searches.get(
+                current_search = await aexa.websets.searches.get(
                     webset_id=input_data.webset_id, id=search_id
                 )
                 current_status = (
@@ -442,7 +442,7 @@ class ExaGetWebsetSearchBlock(Block):
         # Use AsyncExa SDK
         aexa = AsyncExa(api_key=credentials.api_key.get_secret_value())
 
-        sdk_search = aexa.websets.searches.get(
+        sdk_search = await aexa.websets.searches.get(
             webset_id=input_data.webset_id, id=input_data.search_id
         )
 
@@ -523,7 +523,7 @@ class ExaCancelWebsetSearchBlock(Block):
         # Use AsyncExa SDK
         aexa = AsyncExa(api_key=credentials.api_key.get_secret_value())
 
-        canceled_search = aexa.websets.searches.cancel(
+        canceled_search = await aexa.websets.searches.cancel(
             webset_id=input_data.webset_id, id=input_data.search_id
         )
 
@@ -604,7 +604,7 @@ class ExaFindOrCreateSearchBlock(Block):
         aexa = AsyncExa(api_key=credentials.api_key.get_secret_value())
 
         # Get webset to check existing searches
-        webset = aexa.websets.get(id=input_data.webset_id)
+        webset = await aexa.websets.get(id=input_data.webset_id)
 
         # Look for existing search with same query
         existing_search = None
@@ -636,7 +636,7 @@ class ExaFindOrCreateSearchBlock(Block):
             if input_data.entity_type != SearchEntityType.AUTO:
                 payload["entity"] = {"type": input_data.entity_type.value}
 
-            sdk_search = aexa.websets.searches.create(
+            sdk_search = await aexa.websets.searches.create(
                 webset_id=input_data.webset_id, params=payload
             )
 

autogpt_platform/backend/backend/blocks/fal/ai_video_generator.py

@@ -17,8 +17,11 @@ from backend.data.block import (
     BlockSchemaInput,
     BlockSchemaOutput,
 )
+from backend.data.execution import ExecutionContext
 from backend.data.model import SchemaField
+from backend.util.file import store_media_file
 from backend.util.request import ClientResponseError, Requests
+from backend.util.type import MediaFileType
 
 logger = logging.getLogger(__name__)
 
@@ -64,9 +67,13 @@ class AIVideoGeneratorBlock(Block):
                 "credentials": TEST_CREDENTIALS_INPUT,
             },
             test_credentials=TEST_CREDENTIALS,
-            test_output=[("video_url", "https://fal.media/files/example/video.mp4")],
+            test_output=[
+                # Output will be a workspace ref or data URI depending on context
+                ("video_url", lambda x: x.startswith(("workspace://", "data:"))),
+            ],
             test_mock={
-                "generate_video": lambda *args, **kwargs: "https://fal.media/files/example/video.mp4"
+                # Use data URI to avoid HTTP requests during tests
+                "generate_video": lambda *args, **kwargs: "data:video/mp4;base64,AAAA"
             },
         )
 
@@ -208,11 +215,22 @@ class AIVideoGeneratorBlock(Block):
             raise RuntimeError(f"API request failed: {str(e)}")
 
     async def run(
-        self, input_data: Input, *, credentials: FalCredentials, **kwargs
+        self,
+        input_data: Input,
+        *,
+        credentials: FalCredentials,
+        execution_context: ExecutionContext,
+        **kwargs,
     ) -> BlockOutput:
         try:
             video_url = await self.generate_video(input_data, credentials)
-            yield "video_url", video_url
+            # Store the generated video to the user's workspace for persistence
+            stored_url = await store_media_file(
+                file=MediaFileType(video_url),
+                execution_context=execution_context,
+                return_format="for_block_output",
+            )
+            yield "video_url", stored_url
         except Exception as e:
             error_message = str(e)
             yield "error", error_message

autogpt_platform/backend/backend/blocks/flux_kontext.py

@@ -12,6 +12,7 @@ from backend.data.block import (
     BlockSchemaInput,
     BlockSchemaOutput,
 )
+from backend.data.execution import ExecutionContext
 from backend.data.model import (
     APIKeyCredentials,
     CredentialsField,
@@ -121,10 +122,12 @@ class AIImageEditorBlock(Block):
                 "credentials": TEST_CREDENTIALS_INPUT,
             },
             test_output=[
-                ("output_image", "https://replicate.com/output/edited-image.png"),
+                # Output will be a workspace ref or data URI depending on context
+                ("output_image", lambda x: x.startswith(("workspace://", "data:"))),
             ],
             test_mock={
-                "run_model": lambda *args, **kwargs: "https://replicate.com/output/edited-image.png",
+                # Use data URI to avoid HTTP requests during tests
+                "run_model": lambda *args, **kwargs: "data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAAEAAAABCAYAAAAfFcSJAAAADUlEQVR42mNk+M9QDwADhgGAWjR9awAAAABJRU5ErkJggg==",
             },
             test_credentials=TEST_CREDENTIALS,
         )
@@ -134,8 +137,7 @@ class AIImageEditorBlock(Block):
         input_data: Input,
         *,
         credentials: APIKeyCredentials,
-        graph_exec_id: str,
-        user_id: str,
+        execution_context: ExecutionContext,
         **kwargs,
     ) -> BlockOutput:
         result = await self.run_model(
@@ -144,20 +146,25 @@ class AIImageEditorBlock(Block):
             prompt=input_data.prompt,
             input_image_b64=(
                 await store_media_file(
-                    graph_exec_id=graph_exec_id,
                     file=input_data.input_image,
-                    user_id=user_id,
-                    return_content=True,
+                    execution_context=execution_context,
+                    return_format="for_external_api",  # Get content for Replicate API
                 )
                 if input_data.input_image
                 else None
             ),
             aspect_ratio=input_data.aspect_ratio.value,
             seed=input_data.seed,
-            user_id=user_id,
-            graph_exec_id=graph_exec_id,
+            user_id=execution_context.user_id or "",
+            graph_exec_id=execution_context.graph_exec_id or "",
         )
-        yield "output_image", result
+        # Store the generated image to the user's workspace for persistence
+        stored_url = await store_media_file(
+            file=result,
+            execution_context=execution_context,
+            return_format="for_block_output",
+        )
+        yield "output_image", stored_url
 
     async def run_model(
         self,

autogpt_platform/backend/backend/blocks/google/gmail.py

@@ -21,6 +21,7 @@ from backend.data.block import (
     BlockSchemaInput,
     BlockSchemaOutput,
 )
+from backend.data.execution import ExecutionContext
 from backend.data.model import SchemaField
 from backend.util.file import MediaFileType, get_exec_file_path, store_media_file
 from backend.util.settings import Settings
@@ -95,8 +96,7 @@ def _make_mime_text(
 
 async def create_mime_message(
     input_data,
-    graph_exec_id: str,
-    user_id: str,
+    execution_context: ExecutionContext,
 ) -> str:
     """Create a MIME message with attachments and return base64-encoded raw message."""
 
@@ -117,12 +117,12 @@ async def create_mime_message(
     if input_data.attachments:
         for attach in input_data.attachments:
             local_path = await store_media_file(
-                user_id=user_id,
-                graph_exec_id=graph_exec_id,
                 file=attach,
-                return_content=False,
+                execution_context=execution_context,
+                return_format="for_local_processing",
             )
-            abs_path = get_exec_file_path(graph_exec_id, local_path)
+            assert execution_context.graph_exec_id  # Validated by store_media_file
+            abs_path = get_exec_file_path(execution_context.graph_exec_id, local_path)
             part = MIMEBase("application", "octet-stream")
             with open(abs_path, "rb") as f:
                 part.set_payload(f.read())
@@ -582,27 +582,25 @@ class GmailSendBlock(GmailBase):
         input_data: Input,
         *,
         credentials: GoogleCredentials,
-        graph_exec_id: str,
-        user_id: str,
+        execution_context: ExecutionContext,
         **kwargs,
     ) -> BlockOutput:
         service = self._build_service(credentials, **kwargs)
         result = await self._send_email(
             service,
             input_data,
-            graph_exec_id,
-            user_id,
+            execution_context,
         )
         yield "result", result
 
     async def _send_email(
-        self, service, input_data: Input, graph_exec_id: str, user_id: str
+        self, service, input_data: Input, execution_context: ExecutionContext
     ) -> dict:
         if not input_data.to or not input_data.subject or not input_data.body:
             raise ValueError(
                 "At least one recipient, subject, and body are required for sending an email"
             )
-        raw_message = await create_mime_message(input_data, graph_exec_id, user_id)
+        raw_message = await create_mime_message(input_data, execution_context)
         sent_message = await asyncio.to_thread(
             lambda: service.users()
             .messages()
@@ -692,30 +690,28 @@ class GmailCreateDraftBlock(GmailBase):
         input_data: Input,
         *,
         credentials: GoogleCredentials,
-        graph_exec_id: str,
-        user_id: str,
+        execution_context: ExecutionContext,
         **kwargs,
     ) -> BlockOutput:
         service = self._build_service(credentials, **kwargs)
         result = await self._create_draft(
             service,
             input_data,
-            graph_exec_id,
-            user_id,
+            execution_context,
         )
         yield "result", GmailDraftResult(
             id=result["id"], message_id=result["message"]["id"], status="draft_created"
         )
 
     async def _create_draft(
-        self, service, input_data: Input, graph_exec_id: str, user_id: str
+        self, service, input_data: Input, execution_context: ExecutionContext
     ) -> dict:
         if not input_data.to or not input_data.subject:
             raise ValueError(
                 "At least one recipient and subject are required for creating a draft"
             )
 
-        raw_message = await create_mime_message(input_data, graph_exec_id, user_id)
+        raw_message = await create_mime_message(input_data, execution_context)
         draft = await asyncio.to_thread(
             lambda: service.users()
             .drafts()
@@ -1100,7 +1096,7 @@ class GmailGetThreadBlock(GmailBase):
 
 
 async def _build_reply_message(
-    service, input_data, graph_exec_id: str, user_id: str
+    service, input_data, execution_context: ExecutionContext
 ) -> tuple[str, str]:
     """
     Builds a reply MIME message for Gmail threads.
@@ -1190,12 +1186,12 @@ async def _build_reply_message(
     # Handle attachments
     for attach in input_data.attachments:
         local_path = await store_media_file(
-            user_id=user_id,
-            graph_exec_id=graph_exec_id,
             file=attach,
-            return_content=False,
+            execution_context=execution_context,
+            return_format="for_local_processing",
         )
-        abs_path = get_exec_file_path(graph_exec_id, local_path)
+        assert execution_context.graph_exec_id  # Validated by store_media_file
+        abs_path = get_exec_file_path(execution_context.graph_exec_id, local_path)
         part = MIMEBase("application", "octet-stream")
         with open(abs_path, "rb") as f:
             part.set_payload(f.read())
@@ -1311,16 +1307,14 @@ class GmailReplyBlock(GmailBase):
         input_data: Input,
         *,
         credentials: GoogleCredentials,
-        graph_exec_id: str,
-        user_id: str,
+        execution_context: ExecutionContext,
         **kwargs,
     ) -> BlockOutput:
         service = self._build_service(credentials, **kwargs)
         message = await self._reply(
             service,
             input_data,
-            graph_exec_id,
-            user_id,
+            execution_context,
         )
         yield "messageId", message["id"]
         yield "threadId", message.get("threadId", input_data.threadId)
@@ -1343,11 +1337,11 @@ class GmailReplyBlock(GmailBase):
         yield "email", email
 
     async def _reply(
-        self, service, input_data: Input, graph_exec_id: str, user_id: str
+        self, service, input_data: Input, execution_context: ExecutionContext
     ) -> dict:
         # Build the reply message using the shared helper
         raw, thread_id = await _build_reply_message(
-            service, input_data, graph_exec_id, user_id
+            service, input_data, execution_context
         )
 
         # Send the message
@@ -1441,16 +1435,14 @@ class GmailDraftReplyBlock(GmailBase):
         input_data: Input,
         *,
         credentials: GoogleCredentials,
-        graph_exec_id: str,
-        user_id: str,
+        execution_context: ExecutionContext,
         **kwargs,
     ) -> BlockOutput:
         service = self._build_service(credentials, **kwargs)
         draft = await self._create_draft_reply(
             service,
             input_data,
-            graph_exec_id,
-            user_id,
+            execution_context,
         )
         yield "draftId", draft["id"]
         yield "messageId", draft["message"]["id"]
@@ -1458,11 +1450,11 @@ class GmailDraftReplyBlock(GmailBase):
         yield "status", "draft_created"
 
     async def _create_draft_reply(
-        self, service, input_data: Input, graph_exec_id: str, user_id: str
+        self, service, input_data: Input, execution_context: ExecutionContext
     ) -> dict:
         # Build the reply message using the shared helper
         raw, thread_id = await _build_reply_message(
-            service, input_data, graph_exec_id, user_id
+            service, input_data, execution_context
         )
 
         # Create draft with proper thread association
@@ -1629,23 +1621,21 @@ class GmailForwardBlock(GmailBase):
         input_data: Input,
         *,
         credentials: GoogleCredentials,
-        graph_exec_id: str,
-        user_id: str,
+        execution_context: ExecutionContext,
         **kwargs,
     ) -> BlockOutput:
         service = self._build_service(credentials, **kwargs)
         result = await self._forward_message(
             service,
             input_data,
-            graph_exec_id,
-            user_id,
+            execution_context,
         )
         yield "messageId", result["id"]
         yield "threadId", result.get("threadId", "")
         yield "status", "forwarded"
 
     async def _forward_message(
-        self, service, input_data: Input, graph_exec_id: str, user_id: str
+        self, service, input_data: Input, execution_context: ExecutionContext
     ) -> dict:
         if not input_data.to:
             raise ValueError("At least one recipient is required for forwarding")
@@ -1727,12 +1717,12 @@ To: {original_to}
         # Add any additional attachments
         for attach in input_data.additionalAttachments:
             local_path = await store_media_file(
-                user_id=user_id,
-                graph_exec_id=graph_exec_id,
                 file=attach,
-                return_content=False,
+                execution_context=execution_context,
+                return_format="for_local_processing",
             )
-            abs_path = get_exec_file_path(graph_exec_id, local_path)
+            assert execution_context.graph_exec_id  # Validated by store_media_file
+            abs_path = get_exec_file_path(execution_context.graph_exec_id, local_path)
             part = MIMEBase("application", "octet-stream")
             with open(abs_path, "rb") as f:
                 part.set_payload(f.read())

autogpt_platform/backend/backend/blocks/helpers/review.py

@@ -9,7 +9,7 @@ from typing import Any, Optional
 from prisma.enums import ReviewStatus
 from pydantic import BaseModel
 
-from backend.data.execution import ExecutionContext, ExecutionStatus
+from backend.data.execution import ExecutionStatus
 from backend.data.human_review import ReviewResult
 from backend.executor.manager import async_update_node_execution_status
 from backend.util.clients import get_database_manager_async_client
@@ -28,6 +28,11 @@ class ReviewDecision(BaseModel):
 class HITLReviewHelper:
     """Helper class for Human-In-The-Loop review operations."""
 
+    @staticmethod
+    async def check_approval(**kwargs) -> Optional[ReviewResult]:
+        """Check if there's an existing approval for this node execution."""
+        return await get_database_manager_async_client().check_approval(**kwargs)
+
     @staticmethod
     async def get_or_create_human_review(**kwargs) -> Optional[ReviewResult]:
         """Create or retrieve a human review from the database."""
@@ -55,11 +60,11 @@ class HITLReviewHelper:
     async def _handle_review_request(
         input_data: Any,
         user_id: str,
+        node_id: str,
         node_exec_id: str,
         graph_exec_id: str,
         graph_id: str,
         graph_version: int,
-        execution_context: ExecutionContext,
         block_name: str = "Block",
         editable: bool = False,
     ) -> Optional[ReviewResult]:
@@ -69,11 +74,11 @@ class HITLReviewHelper:
         Args:
             input_data: The input data to be reviewed
             user_id: ID of the user requesting the review
+            node_id: ID of the node in the graph definition
             node_exec_id: ID of the node execution
             graph_exec_id: ID of the graph execution
             graph_id: ID of the graph
             graph_version: Version of the graph
-            execution_context: Current execution context
             block_name: Name of the block requesting review
             editable: Whether the reviewer can edit the data
 
@@ -83,15 +88,41 @@ class HITLReviewHelper:
         Raises:
             Exception: If review creation or status update fails
         """
-        # Skip review if safe mode is disabled - return auto-approved result
-        if not execution_context.human_in_the_loop_safe_mode:
+        # Note: Safe mode checks (human_in_the_loop_safe_mode, sensitive_action_safe_mode)
+        # are handled by the caller:
+        # - HITL blocks check human_in_the_loop_safe_mode in their run() method
+        # - Sensitive action blocks check sensitive_action_safe_mode in is_block_exec_need_review()
+        # This function only handles checking for existing approvals.
+
+        # Check if this node has already been approved (normal or auto-approval)
+        if approval_result := await HITLReviewHelper.check_approval(
+            node_exec_id=node_exec_id,
+            graph_exec_id=graph_exec_id,
+            node_id=node_id,
+            user_id=user_id,
+            input_data=input_data,
+        ):
             logger.info(
-                f"Block {block_name} skipping review for node {node_exec_id} - safe mode disabled"
+                f"Block {block_name} skipping review for node {node_exec_id} - "
+                f"found existing approval"
+            )
+            # Return a new ReviewResult with the current node_exec_id but approved status
+            # For auto-approvals, always use current input_data
+            # For normal approvals, use approval_result.data unless it's None
+            is_auto_approval = approval_result.node_exec_id != node_exec_id
+            approved_data = (
+                input_data
+                if is_auto_approval
+                else (
+                    approval_result.data
+                    if approval_result.data is not None
+                    else input_data
+                )
             )
             return ReviewResult(
-                data=input_data,
+                data=approved_data,
                 status=ReviewStatus.APPROVED,
-                message="Auto-approved (safe mode disabled)",
+                message=approval_result.message,
                 processed=True,
                 node_exec_id=node_exec_id,
             )
@@ -103,7 +134,7 @@ class HITLReviewHelper:
             graph_id=graph_id,
             graph_version=graph_version,
             input_data=input_data,
-            message=f"Review required for {block_name} execution",
+            message=block_name,  # Use block_name directly as the message
             editable=editable,
         )
 
@@ -129,11 +160,11 @@ class HITLReviewHelper:
     async def handle_review_decision(
         input_data: Any,
         user_id: str,
+        node_id: str,
         node_exec_id: str,
         graph_exec_id: str,
         graph_id: str,
         graph_version: int,
-        execution_context: ExecutionContext,
         block_name: str = "Block",
         editable: bool = False,
     ) -> Optional[ReviewDecision]:
@@ -143,11 +174,11 @@ class HITLReviewHelper:
         Args:
             input_data: The input data to be reviewed
             user_id: ID of the user requesting the review
+            node_id: ID of the node in the graph definition
             node_exec_id: ID of the node execution
             graph_exec_id: ID of the graph execution
             graph_id: ID of the graph
             graph_version: Version of the graph
-            execution_context: Current execution context
             block_name: Name of the block requesting review
             editable: Whether the reviewer can edit the data
 
@@ -158,11 +189,11 @@ class HITLReviewHelper:
         review_result = await HITLReviewHelper._handle_review_request(
             input_data=input_data,
             user_id=user_id,
+            node_id=node_id,
             node_exec_id=node_exec_id,
             graph_exec_id=graph_exec_id,
             graph_id=graph_id,
             graph_version=graph_version,
-            execution_context=execution_context,
             block_name=block_name,
             editable=editable,
         )

autogpt_platform/backend/backend/blocks/http.py

@@ -15,6 +15,7 @@ from backend.data.block import (
     BlockSchemaInput,
     BlockSchemaOutput,
 )
+from backend.data.execution import ExecutionContext
 from backend.data.model import (
     CredentialsField,
     CredentialsMetaInput,
@@ -116,10 +117,9 @@ class SendWebRequestBlock(Block):
 
     @staticmethod
     async def _prepare_files(
-        graph_exec_id: str,
+        execution_context: ExecutionContext,
         files_name: str,
         files: list[MediaFileType],
-        user_id: str,
     ) -> list[tuple[str, tuple[str, BytesIO, str]]]:
         """
         Prepare files for the request by storing them and reading their content.
@@ -127,11 +127,16 @@ class SendWebRequestBlock(Block):
         (files_name, (filename, BytesIO, mime_type))
         """
         files_payload: list[tuple[str, tuple[str, BytesIO, str]]] = []
+        graph_exec_id = execution_context.graph_exec_id
+        if graph_exec_id is None:
+            raise ValueError("graph_exec_id is required for file operations")
 
         for media in files:
             # Normalise to a list so we can repeat the same key
             rel_path = await store_media_file(
-                graph_exec_id, media, user_id, return_content=False
+                file=media,
+                execution_context=execution_context,
+                return_format="for_local_processing",
             )
             abs_path = get_exec_file_path(graph_exec_id, rel_path)
             async with aiofiles.open(abs_path, "rb") as f:
@@ -143,7 +148,7 @@ class SendWebRequestBlock(Block):
         return files_payload
 
     async def run(
-        self, input_data: Input, *, graph_exec_id: str, user_id: str, **kwargs
+        self, input_data: Input, *, execution_context: ExecutionContext, **kwargs
     ) -> BlockOutput:
         # ─── Parse/normalise body ────────────────────────────────────
         body = input_data.body
@@ -174,7 +179,7 @@ class SendWebRequestBlock(Block):
         files_payload: list[tuple[str, tuple[str, BytesIO, str]]] = []
         if use_files:
             files_payload = await self._prepare_files(
-                graph_exec_id, input_data.files_name, input_data.files, user_id
+                execution_context, input_data.files_name, input_data.files
             )
 
         # Enforce body format rules
@@ -238,9 +243,8 @@ class SendAuthenticatedWebRequestBlock(SendWebRequestBlock):
         self,
         input_data: Input,
         *,
-        graph_exec_id: str,
+        execution_context: ExecutionContext,
         credentials: HostScopedCredentials,
-        user_id: str,
         **kwargs,
     ) -> BlockOutput:
         # Create SendWebRequestBlock.Input from our input (removing credentials field)
@@ -271,6 +275,6 @@ class SendAuthenticatedWebRequestBlock(SendWebRequestBlock):
 
         # Use parent class run method
         async for output_name, output_data in super().run(
-            base_input, graph_exec_id=graph_exec_id, user_id=user_id, **kwargs
+            base_input, execution_context=execution_context, **kwargs
         ):
             yield output_name, output_data

autogpt_platform/backend/backend/blocks/human_in_the_loop.py

@@ -97,6 +97,7 @@ class HumanInTheLoopBlock(Block):
         input_data: Input,
         *,
         user_id: str,
+        node_id: str,
         node_exec_id: str,
         graph_exec_id: str,
         graph_id: str,
@@ -115,12 +116,12 @@ class HumanInTheLoopBlock(Block):
         decision = await self.handle_review_decision(
             input_data=input_data.data,
             user_id=user_id,
+            node_id=node_id,
             node_exec_id=node_exec_id,
             graph_exec_id=graph_exec_id,
             graph_id=graph_id,
             graph_version=graph_version,
-            execution_context=execution_context,
-            block_name=self.name,
+            block_name=input_data.name,  # Use user-provided name instead of block type
             editable=input_data.editable,
         )
 

autogpt_platform/backend/backend/blocks/io.py

@@ -12,6 +12,7 @@ from backend.data.block import (
     BlockSchemaInput,
     BlockType,
 )
+from backend.data.execution import ExecutionContext
 from backend.data.model import SchemaField
 from backend.util.file import store_media_file
 from backend.util.mock import MockObject
@@ -462,18 +463,21 @@ class AgentFileInputBlock(AgentInputBlock):
         self,
         input_data: Input,
         *,
-        graph_exec_id: str,
-        user_id: str,
+        execution_context: ExecutionContext,
         **kwargs,
     ) -> BlockOutput:
         if not input_data.value:
             return
 
+        # Determine return format based on user preference
+        # for_external_api: always returns data URI (base64) - honors "Produce Base64 Output"
+        # for_block_output: smart format - workspace:// in CoPilot, data URI in graphs
+        return_format = "for_external_api" if input_data.base_64 else "for_block_output"
+
         yield "result", await store_media_file(
-            graph_exec_id=graph_exec_id,
             file=input_data.value,
-            user_id=user_id,
-            return_content=input_data.base_64,
+            execution_context=execution_context,
+            return_format=return_format,
         )
 
 

autogpt_platform/backend/backend/blocks/linear/_api.py

@@ -162,8 +162,16 @@ class LinearClient:
                 "searchTerm": team_name,
             }
 
-            team_id = await self.query(query, variables)
-            return team_id["teams"]["nodes"][0]["id"]
+            result = await self.query(query, variables)
+            nodes = result["teams"]["nodes"]
+
+            if not nodes:
+                raise LinearAPIException(
+                    f"Team '{team_name}' not found. Check the team name or key and try again.",
+                    status_code=404,
+                )
+
+            return nodes[0]["id"]
         except LinearAPIException as e:
             raise e
 
@@ -240,17 +248,44 @@ class LinearClient:
         except LinearAPIException as e:
             raise e
 
-    async def try_search_issues(self, term: str) -> list[Issue]:
+    async def try_search_issues(
+        self,
+        term: str,
+        max_results: int = 10,
+        team_id: str | None = None,
+    ) -> list[Issue]:
         try:
             query = """
-                query SearchIssues($term: String!, $includeComments: Boolean!) {
-                    searchIssues(term: $term, includeComments: $includeComments) {
+                query SearchIssues(
+                    $term: String!,
+                    $first: Int,
+                    $teamId: String
+                ) {
+                    searchIssues(
+                        term: $term,
+                        first: $first,
+                        teamId: $teamId
+                    ) {
                         nodes {
                             id
                             identifier
                             title
                             description
                             priority
+                            createdAt
+                            state {
+                                id
+                                name
+                                type
+                            }
+                            project {
+                                id
+                                name
+                            }
+                            assignee {
+                                id
+                                name
+                            }
                         }
                     }
                 }
@@ -258,7 +293,8 @@ class LinearClient:
 
             variables: dict[str, Any] = {
                 "term": term,
-                "includeComments": True,
+                "first": max_results,
+                "teamId": team_id,
             }
 
             issues = await self.query(query, variables)

autogpt_platform/backend/backend/blocks/linear/issues.py

@@ -17,7 +17,7 @@ from ._config import (
     LinearScope,
     linear,
 )
-from .models import CreateIssueResponse, Issue
+from .models import CreateIssueResponse, Issue, State
 
 
 class LinearCreateIssueBlock(Block):
@@ -135,9 +135,20 @@ class LinearSearchIssuesBlock(Block):
             description="Linear credentials with read permissions",
             required_scopes={LinearScope.READ},
         )
+        max_results: int = SchemaField(
+            description="Maximum number of results to return",
+            default=10,
+            ge=1,
+            le=100,
+        )
+        team_name: str | None = SchemaField(
+            description="Optional team name to filter results (e.g., 'Internal', 'Open Source')",
+            default=None,
+        )
 
     class Output(BlockSchemaOutput):
         issues: list[Issue] = SchemaField(description="List of issues")
+        error: str = SchemaField(description="Error message if the search failed")
 
     def __init__(self):
         super().__init__(
@@ -145,8 +156,11 @@ class LinearSearchIssuesBlock(Block):
             description="Searches for issues on Linear",
             input_schema=self.Input,
             output_schema=self.Output,
+            categories={BlockCategory.PRODUCTIVITY, BlockCategory.ISSUE_TRACKING},
             test_input={
                 "term": "Test issue",
+                "max_results": 10,
+                "team_name": None,
                 "credentials": TEST_CREDENTIALS_INPUT_OAUTH,
             },
             test_credentials=TEST_CREDENTIALS_OAUTH,
@@ -156,10 +170,14 @@ class LinearSearchIssuesBlock(Block):
                     [
                         Issue(
                             id="abc123",
-                            identifier="abc123",
+                            identifier="TST-123",
                             title="Test issue",
                             description="Test description",
                             priority=1,
+                            state=State(
+                                id="state1", name="In Progress", type="started"
+                            ),
+                            createdAt="2026-01-15T10:00:00.000Z",
                         )
                     ],
                 )
@@ -168,10 +186,12 @@ class LinearSearchIssuesBlock(Block):
                 "search_issues": lambda *args, **kwargs: [
                     Issue(
                         id="abc123",
-                        identifier="abc123",
+                        identifier="TST-123",
                         title="Test issue",
                         description="Test description",
                         priority=1,
+                        state=State(id="state1", name="In Progress", type="started"),
+                        createdAt="2026-01-15T10:00:00.000Z",
                     )
                 ]
             },
@@ -181,10 +201,22 @@ class LinearSearchIssuesBlock(Block):
     async def search_issues(
         credentials: OAuth2Credentials | APIKeyCredentials,
         term: str,
+        max_results: int = 10,
+        team_name: str | None = None,
     ) -> list[Issue]:
         client = LinearClient(credentials=credentials)
-        response: list[Issue] = await client.try_search_issues(term=term)
-        return response
+
+        # Resolve team name to ID if provided
+        # Raises LinearAPIException with descriptive message if team not found
+        team_id: str | None = None
+        if team_name:
+            team_id = await client.try_get_team_by_name(team_name=team_name)
+
+        return await client.try_search_issues(
+            term=term,
+            max_results=max_results,
+            team_id=team_id,
+        )
 
     async def run(
         self,
@@ -196,7 +228,10 @@ class LinearSearchIssuesBlock(Block):
         """Execute the issue search"""
         try:
             issues = await self.search_issues(
-                credentials=credentials, term=input_data.term
+                credentials=credentials,
+                term=input_data.term,
+                max_results=input_data.max_results,
+                team_name=input_data.team_name,
             )
             yield "issues", issues
         except LinearAPIException as e:

autogpt_platform/backend/backend/blocks/linear/models.py

@@ -36,12 +36,21 @@ class Project(BaseModel):
     content: str | None = None
 
 
+class State(BaseModel):
+    id: str
+    name: str
+    type: str | None = (
+        None  # Workflow state type (e.g., "triage", "backlog", "started", "completed", "canceled")
+    )
+
+
 class Issue(BaseModel):
     id: str
     identifier: str
     title: str
     description: str | None
     priority: int
+    state: State | None = None
     project: Project | None = None
     createdAt: str | None = None
     comments: list[Comment] | None = None

autogpt_platform/backend/backend/blocks/llm.py

@@ -32,7 +32,7 @@ from backend.data.model import (
 from backend.integrations.providers import ProviderName
 from backend.util import json
 from backend.util.logging import TruncatedLogger
-from backend.util.prompt import compress_prompt, estimate_token_count
+from backend.util.prompt import compress_context, estimate_token_count
 from backend.util.text import TextFormatter
 
 logger = TruncatedLogger(logging.getLogger(__name__), "[LLM-Block]")
@@ -115,7 +115,7 @@ class LlmModel(str, Enum, metaclass=LlmModelMeta):
     CLAUDE_4_5_OPUS = "claude-opus-4-5-20251101"
     CLAUDE_4_5_SONNET = "claude-sonnet-4-5-20250929"
     CLAUDE_4_5_HAIKU = "claude-haiku-4-5-20251001"
-    CLAUDE_3_7_SONNET = "claude-3-7-sonnet-20250219"
+    CLAUDE_4_6_OPUS = "claude-opus-4-6"
     CLAUDE_3_HAIKU = "claude-3-haiku-20240307"
     # AI/ML API models
     AIML_API_QWEN2_5_72B = "Qwen/Qwen2.5-72B-Instruct-Turbo"
@@ -271,6 +271,9 @@ MODEL_METADATA = {
     LlmModel.CLAUDE_4_SONNET: ModelMetadata(
         "anthropic", 200000, 64000, "Claude Sonnet 4", "Anthropic", "Anthropic", 2
     ),  # claude-4-sonnet-20250514
+    LlmModel.CLAUDE_4_6_OPUS: ModelMetadata(
+        "anthropic", 200000, 128000, "Claude Opus 4.6", "Anthropic", "Anthropic", 3
+    ),  # claude-opus-4-6
     LlmModel.CLAUDE_4_5_OPUS: ModelMetadata(
         "anthropic", 200000, 64000, "Claude Opus 4.5", "Anthropic", "Anthropic", 3
     ),  # claude-opus-4-5-20251101
@@ -280,9 +283,6 @@ MODEL_METADATA = {
     LlmModel.CLAUDE_4_5_HAIKU: ModelMetadata(
         "anthropic", 200000, 64000, "Claude Haiku 4.5", "Anthropic", "Anthropic", 2
     ),  # claude-haiku-4-5-20251001
-    LlmModel.CLAUDE_3_7_SONNET: ModelMetadata(
-        "anthropic", 200000, 64000, "Claude 3.7 Sonnet", "Anthropic", "Anthropic", 2
-    ),  # claude-3-7-sonnet-20250219
     LlmModel.CLAUDE_3_HAIKU: ModelMetadata(
         "anthropic", 200000, 4096, "Claude 3 Haiku", "Anthropic", "Anthropic", 1
     ),  # claude-3-haiku-20240307
@@ -531,12 +531,12 @@ class LLMResponse(BaseModel):
 
 def convert_openai_tool_fmt_to_anthropic(
     openai_tools: list[dict] | None = None,
-) -> Iterable[ToolParam] | anthropic.NotGiven:
+) -> Iterable[ToolParam] | anthropic.Omit:
     """
     Convert OpenAI tool format to Anthropic tool format.
     """
     if not openai_tools or len(openai_tools) == 0:
-        return anthropic.NOT_GIVEN
+        return anthropic.omit
 
     anthropic_tools = []
     for tool in openai_tools:
@@ -596,10 +596,10 @@ def extract_openai_tool_calls(response) -> list[ToolContentBlock] | None:
 
 def get_parallel_tool_calls_param(
     llm_model: LlmModel, parallel_tool_calls: bool | None
-):
+) -> bool | openai.Omit:
     """Get the appropriate parallel_tool_calls parameter for OpenAI-compatible APIs."""
     if llm_model.startswith("o") or parallel_tool_calls is None:
-        return openai.NOT_GIVEN
+        return openai.omit
     return parallel_tool_calls
 
 
@@ -638,11 +638,18 @@ async def llm_call(
     context_window = llm_model.context_window
 
     if compress_prompt_to_fit:
-        prompt = compress_prompt(
+        result = await compress_context(
             messages=prompt,
             target_tokens=llm_model.context_window // 2,
-            lossy_ok=True,
+            client=None,  # Truncation-only, no LLM summarization
+            reserve=0,  # Caller handles response token budget separately
         )
+        if result.error:
+            logger.warning(
+                f"Prompt compression did not meet target: {result.error}. "
+                f"Proceeding with {result.token_count} tokens."
+            )
+        prompt = result.messages
 
     # Calculate available tokens based on context window and input length
     estimated_input_tokens = estimate_token_count(prompt)

autogpt_platform/backend/backend/blocks/media.py

@@ -1,251 +0,0 @@
-import os
-import tempfile
-from typing import Literal, Optional
-
-from moviepy.audio.io.AudioFileClip import AudioFileClip
-from moviepy.video.fx.Loop import Loop
-from moviepy.video.io.VideoFileClip import VideoFileClip
-
-from backend.data.block import (
-    Block,
-    BlockCategory,
-    BlockOutput,
-    BlockSchemaInput,
-    BlockSchemaOutput,
-)
-from backend.data.model import SchemaField
-from backend.util.file import MediaFileType, get_exec_file_path, store_media_file
-
-
-class MediaDurationBlock(Block):
-
-    class Input(BlockSchemaInput):
-        media_in: MediaFileType = SchemaField(
-            description="Media input (URL, data URI, or local path)."
-        )
-        is_video: bool = SchemaField(
-            description="Whether the media is a video (True) or audio (False).",
-            default=True,
-        )
-
-    class Output(BlockSchemaOutput):
-        duration: float = SchemaField(
-            description="Duration of the media file (in seconds)."
-        )
-
-    def __init__(self):
-        super().__init__(
-            id="d8b91fd4-da26-42d4-8ecb-8b196c6d84b6",
-            description="Block to get the duration of a media file.",
-            categories={BlockCategory.MULTIMEDIA},
-            input_schema=MediaDurationBlock.Input,
-            output_schema=MediaDurationBlock.Output,
-        )
-
-    async def run(
-        self,
-        input_data: Input,
-        *,
-        graph_exec_id: str,
-        user_id: str,
-        **kwargs,
-    ) -> BlockOutput:
-        # 1) Store the input media locally
-        local_media_path = await store_media_file(
-            graph_exec_id=graph_exec_id,
-            file=input_data.media_in,
-            user_id=user_id,
-            return_content=False,
-        )
-        media_abspath = get_exec_file_path(graph_exec_id, local_media_path)
-
-        # 2) Load the clip
-        if input_data.is_video:
-            clip = VideoFileClip(media_abspath)
-        else:
-            clip = AudioFileClip(media_abspath)
-
-        yield "duration", clip.duration
-
-
-class LoopVideoBlock(Block):
-    """
-    Block for looping (repeating) a video clip until a given duration or number of loops.
-    """
-
-    class Input(BlockSchemaInput):
-        video_in: MediaFileType = SchemaField(
-            description="The input video (can be a URL, data URI, or local path)."
-        )
-        # Provide EITHER a `duration` or `n_loops` or both. We'll demonstrate `duration`.
-        duration: Optional[float] = SchemaField(
-            description="Target duration (in seconds) to loop the video to. If omitted, defaults to no looping.",
-            default=None,
-            ge=0.0,
-        )
-        n_loops: Optional[int] = SchemaField(
-            description="Number of times to repeat the video. If omitted, defaults to 1 (no repeat).",
-            default=None,
-            ge=1,
-        )
-        output_return_type: Literal["file_path", "data_uri"] = SchemaField(
-            description="How to return the output video. Either a relative path or base64 data URI.",
-            default="file_path",
-        )
-
-    class Output(BlockSchemaOutput):
-        video_out: str = SchemaField(
-            description="Looped video returned either as a relative path or a data URI."
-        )
-
-    def __init__(self):
-        super().__init__(
-            id="8bf9eef6-5451-4213-b265-25306446e94b",
-            description="Block to loop a video to a given duration or number of repeats.",
-            categories={BlockCategory.MULTIMEDIA},
-            input_schema=LoopVideoBlock.Input,
-            output_schema=LoopVideoBlock.Output,
-        )
-
-    async def run(
-        self,
-        input_data: Input,
-        *,
-        node_exec_id: str,
-        graph_exec_id: str,
-        user_id: str,
-        **kwargs,
-    ) -> BlockOutput:
-        # 1) Store the input video locally
-        local_video_path = await store_media_file(
-            graph_exec_id=graph_exec_id,
-            file=input_data.video_in,
-            user_id=user_id,
-            return_content=False,
-        )
-        input_abspath = get_exec_file_path(graph_exec_id, local_video_path)
-
-        # 2) Load the clip
-        clip = VideoFileClip(input_abspath)
-
-        # 3) Apply the loop effect
-        looped_clip = clip
-        if input_data.duration:
-            # Loop until we reach the specified duration
-            looped_clip = looped_clip.with_effects([Loop(duration=input_data.duration)])
-        elif input_data.n_loops:
-            looped_clip = looped_clip.with_effects([Loop(n=input_data.n_loops)])
-        else:
-            raise ValueError("Either 'duration' or 'n_loops' must be provided.")
-
-        assert isinstance(looped_clip, VideoFileClip)
-
-        # 4) Save the looped output
-        output_filename = MediaFileType(
-            f"{node_exec_id}_looped_{os.path.basename(local_video_path)}"
-        )
-        output_abspath = get_exec_file_path(graph_exec_id, output_filename)
-
-        looped_clip = looped_clip.with_audio(clip.audio)
-        looped_clip.write_videofile(output_abspath, codec="libx264", audio_codec="aac")
-
-        # Return as data URI
-        video_out = await store_media_file(
-            graph_exec_id=graph_exec_id,
-            file=output_filename,
-            user_id=user_id,
-            return_content=input_data.output_return_type == "data_uri",
-        )
-
-        yield "video_out", video_out
-
-
-class AddAudioToVideoBlock(Block):
-    """
-    Block that adds (attaches) an audio track to an existing video.
-    Optionally scale the volume of the new track.
-    """
-
-    class Input(BlockSchemaInput):
-        video_in: MediaFileType = SchemaField(
-            description="Video input (URL, data URI, or local path)."
-        )
-        audio_in: MediaFileType = SchemaField(
-            description="Audio input (URL, data URI, or local path)."
-        )
-        volume: float = SchemaField(
-            description="Volume scale for the newly attached audio track (1.0 = original).",
-            default=1.0,
-        )
-        output_return_type: Literal["file_path", "data_uri"] = SchemaField(
-            description="Return the final output as a relative path or base64 data URI.",
-            default="file_path",
-        )
-
-    class Output(BlockSchemaOutput):
-        video_out: MediaFileType = SchemaField(
-            description="Final video (with attached audio), as a path or data URI."
-        )
-
-    def __init__(self):
-        super().__init__(
-            id="3503748d-62b6-4425-91d6-725b064af509",
-            description="Block to attach an audio file to a video file using moviepy.",
-            categories={BlockCategory.MULTIMEDIA},
-            input_schema=AddAudioToVideoBlock.Input,
-            output_schema=AddAudioToVideoBlock.Output,
-        )
-
-    async def run(
-        self,
-        input_data: Input,
-        *,
-        node_exec_id: str,
-        graph_exec_id: str,
-        user_id: str,
-        **kwargs,
-    ) -> BlockOutput:
-        # 1) Store the inputs locally
-        local_video_path = await store_media_file(
-            graph_exec_id=graph_exec_id,
-            file=input_data.video_in,
-            user_id=user_id,
-            return_content=False,
-        )
-        local_audio_path = await store_media_file(
-            graph_exec_id=graph_exec_id,
-            file=input_data.audio_in,
-            user_id=user_id,
-            return_content=False,
-        )
-
-        abs_temp_dir = os.path.join(tempfile.gettempdir(), "exec_file", graph_exec_id)
-        video_abspath = os.path.join(abs_temp_dir, local_video_path)
-        audio_abspath = os.path.join(abs_temp_dir, local_audio_path)
-
-        # 2) Load video + audio with moviepy
-        video_clip = VideoFileClip(video_abspath)
-        audio_clip = AudioFileClip(audio_abspath)
-        # Optionally scale volume
-        if input_data.volume != 1.0:
-            audio_clip = audio_clip.with_volume_scaled(input_data.volume)
-
-        # 3) Attach the new audio track
-        final_clip = video_clip.with_audio(audio_clip)
-
-        # 4) Write to output file
-        output_filename = MediaFileType(
-            f"{node_exec_id}_audio_attached_{os.path.basename(local_video_path)}"
-        )
-        output_abspath = os.path.join(abs_temp_dir, output_filename)
-        final_clip.write_videofile(output_abspath, codec="libx264", audio_codec="aac")
-
-        # 5) Return either path or data URI
-        video_out = await store_media_file(
-            graph_exec_id=graph_exec_id,
-            file=output_filename,
-            user_id=user_id,
-            return_content=input_data.output_return_type == "data_uri",
-        )
-
-        yield "video_out", video_out

autogpt_platform/backend/backend/blocks/screenshotone.py

@@ -11,6 +11,7 @@ from backend.data.block import (
     BlockSchemaInput,
     BlockSchemaOutput,
 )
+from backend.data.execution import ExecutionContext
 from backend.data.model import (
     APIKeyCredentials,
     CredentialsField,
@@ -112,8 +113,7 @@ class ScreenshotWebPageBlock(Block):
     @staticmethod
     async def take_screenshot(
         credentials: APIKeyCredentials,
-        graph_exec_id: str,
-        user_id: str,
+        execution_context: ExecutionContext,
         url: str,
         viewport_width: int,
         viewport_height: int,
@@ -155,12 +155,11 @@ class ScreenshotWebPageBlock(Block):
 
         return {
             "image": await store_media_file(
-                graph_exec_id=graph_exec_id,
                 file=MediaFileType(
                     f"data:image/{format.value};base64,{b64encode(content).decode('utf-8')}"
                 ),
-                user_id=user_id,
-                return_content=True,
+                execution_context=execution_context,
+                return_format="for_block_output",
             )
         }
 
@@ -169,15 +168,13 @@ class ScreenshotWebPageBlock(Block):
         input_data: Input,
         *,
         credentials: APIKeyCredentials,
-        graph_exec_id: str,
-        user_id: str,
+        execution_context: ExecutionContext,
         **kwargs,
     ) -> BlockOutput:
         try:
             screenshot_data = await self.take_screenshot(
                 credentials=credentials,
-                graph_exec_id=graph_exec_id,
-                user_id=user_id,
+                execution_context=execution_context,
                 url=input_data.url,
                 viewport_width=input_data.viewport_width,
                 viewport_height=input_data.viewport_height,

autogpt_platform/backend/backend/blocks/spreadsheet.py

@@ -7,6 +7,7 @@ from backend.data.block import (
     BlockSchemaInput,
     BlockSchemaOutput,
 )
+from backend.data.execution import ExecutionContext
 from backend.data.model import ContributorDetails, SchemaField
 from backend.util.file import get_exec_file_path, store_media_file
 from backend.util.type import MediaFileType
@@ -98,7 +99,7 @@ class ReadSpreadsheetBlock(Block):
         )
 
     async def run(
-        self, input_data: Input, *, graph_exec_id: str, user_id: str, **_kwargs
+        self, input_data: Input, *, execution_context: ExecutionContext, **_kwargs
     ) -> BlockOutput:
         import csv
         from io import StringIO
@@ -106,14 +107,16 @@ class ReadSpreadsheetBlock(Block):
         # Determine data source - prefer file_input if provided, otherwise use contents
         if input_data.file_input:
             stored_file_path = await store_media_file(
-                user_id=user_id,
-                graph_exec_id=graph_exec_id,
                 file=input_data.file_input,
-                return_content=False,
+                execution_context=execution_context,
+                return_format="for_local_processing",
             )
 
             # Get full file path
-            file_path = get_exec_file_path(graph_exec_id, stored_file_path)
+            assert execution_context.graph_exec_id  # Validated by store_media_file
+            file_path = get_exec_file_path(
+                execution_context.graph_exec_id, stored_file_path
+            )
             if not Path(file_path).exists():
                 raise ValueError(f"File does not exist: {file_path}")
 

autogpt_platform/backend/backend/blocks/stagehand/blocks.py

@@ -83,7 +83,7 @@ class StagehandRecommendedLlmModel(str, Enum):
     GPT41_MINI = "gpt-4.1-mini-2025-04-14"
 
     # Anthropic
-    CLAUDE_3_7_SONNET = "claude-3-7-sonnet-20250219"
+    CLAUDE_4_5_SONNET = "claude-sonnet-4-5-20250929"
 
     @property
     def provider_name(self) -> str:
@@ -137,7 +137,7 @@ class StagehandObserveBlock(Block):
         model: StagehandRecommendedLlmModel = SchemaField(
             title="LLM Model",
             description="LLM to use for Stagehand (provider is inferred)",
-            default=StagehandRecommendedLlmModel.CLAUDE_3_7_SONNET,
+            default=StagehandRecommendedLlmModel.CLAUDE_4_5_SONNET,
             advanced=False,
         )
         model_credentials: AICredentials = AICredentialsField()
@@ -182,10 +182,7 @@ class StagehandObserveBlock(Block):
         **kwargs,
     ) -> BlockOutput:
 
-        logger.info(f"OBSERVE: Stagehand credentials: {stagehand_credentials}")
-        logger.info(
-            f"OBSERVE: Model credentials: {model_credentials} for provider {model_credentials.provider} secret: {model_credentials.api_key.get_secret_value()}"
-        )
+        logger.debug(f"OBSERVE: Using model provider {model_credentials.provider}")
 
         with disable_signal_handling():
             stagehand = Stagehand(
@@ -230,7 +227,7 @@ class StagehandActBlock(Block):
         model: StagehandRecommendedLlmModel = SchemaField(
             title="LLM Model",
             description="LLM to use for Stagehand (provider is inferred)",
-            default=StagehandRecommendedLlmModel.CLAUDE_3_7_SONNET,
+            default=StagehandRecommendedLlmModel.CLAUDE_4_5_SONNET,
             advanced=False,
         )
         model_credentials: AICredentials = AICredentialsField()
@@ -282,10 +279,7 @@ class StagehandActBlock(Block):
         **kwargs,
     ) -> BlockOutput:
 
-        logger.info(f"ACT: Stagehand credentials: {stagehand_credentials}")
-        logger.info(
-            f"ACT: Model credentials: {model_credentials} for provider {model_credentials.provider} secret: {model_credentials.api_key.get_secret_value()}"
-        )
+        logger.debug(f"ACT: Using model provider {model_credentials.provider}")
 
         with disable_signal_handling():
             stagehand = Stagehand(
@@ -330,7 +324,7 @@ class StagehandExtractBlock(Block):
         model: StagehandRecommendedLlmModel = SchemaField(
             title="LLM Model",
             description="LLM to use for Stagehand (provider is inferred)",
-            default=StagehandRecommendedLlmModel.CLAUDE_3_7_SONNET,
+            default=StagehandRecommendedLlmModel.CLAUDE_4_5_SONNET,
             advanced=False,
         )
         model_credentials: AICredentials = AICredentialsField()
@@ -370,10 +364,7 @@ class StagehandExtractBlock(Block):
         **kwargs,
     ) -> BlockOutput:
 
-        logger.info(f"EXTRACT: Stagehand credentials: {stagehand_credentials}")
-        logger.info(
-            f"EXTRACT: Model credentials: {model_credentials} for provider {model_credentials.provider} secret: {model_credentials.api_key.get_secret_value()}"
-        )
+        logger.debug(f"EXTRACT: Using model provider {model_credentials.provider}")
 
         with disable_signal_handling():
             stagehand = Stagehand(