fix: determine clarification answered state from conversation

Instead of localStorage, check if there's a user message after the
clarification request in the conversation. This works across browsers
since the conversation is stored in the backend.

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

@@ -1834,6 +1834,11 @@ async def _execute_long_running_tool(
             tool_call_id=tool_call_id,
             result=error_response.model_dump_json(),
         )
+        # Generate LLM continuation so user sees explanation even for errors
+        try:
+            await _generate_llm_continuation(session_id=session_id, user_id=user_id)
+        except Exception as llm_err:
+            logger.warning(f"Failed to generate LLM continuation for error: {llm_err}")
     finally:
         await _mark_operation_completed(tool_call_id)
 

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

@@ -2,30 +2,54 @@
 
 from .core import (
     AgentGeneratorNotConfiguredError,
+    AgentJsonValidationError,
+    AgentSummary,
+    DecompositionResult,
+    DecompositionStep,
+    LibraryAgentSummary,
+    MarketplaceAgentSummary,
     decompose_goal,
+    enrich_library_agents_from_steps,
+    extract_search_terms_from_steps,
+    extract_uuids_from_text,
     generate_agent,
     generate_agent_patch,
     get_agent_as_json,
+    get_all_relevant_agents_for_generation,
+    get_library_agent_by_graph_id,
+    get_library_agent_by_id,
+    get_library_agents_for_generation,
     json_to_graph,
     save_agent_to_library,
+    search_marketplace_agents_for_generation,
 )
 from .errors import get_user_message_for_error
 from .service import health_check as check_external_service_health
 from .service import is_external_service_configured
 
 __all__ = [
-    # Core functions
+    "AgentGeneratorNotConfiguredError",
+    "AgentJsonValidationError",
+    "AgentSummary",
+    "DecompositionResult",
+    "DecompositionStep",
+    "LibraryAgentSummary",
+    "MarketplaceAgentSummary",
+    "check_external_service_health",
     "decompose_goal",
+    "enrich_library_agents_from_steps",
+    "extract_search_terms_from_steps",
+    "extract_uuids_from_text",
     "generate_agent",
     "generate_agent_patch",
-    "save_agent_to_library",
     "get_agent_as_json",
-    "json_to_graph",
+    "get_all_relevant_agents_for_generation",
-    # Exceptions
+    "get_library_agent_by_graph_id",
-    "AgentGeneratorNotConfiguredError",
+    "get_library_agent_by_id",
-    # Service
+    "get_library_agents_for_generation",
-    "is_external_service_configured",
-    "check_external_service_health",
-    # Error handling
     "get_user_message_for_error",
+    "is_external_service_configured",
+    "json_to_graph",
+    "save_agent_to_library",
+    "search_marketplace_agents_for_generation",
 ]

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

@@ -1,11 +1,21 @@
 """Core agent generation functions."""
 
 import logging
+import re
 import uuid
-from typing import Any
+from typing import Any, NotRequired, TypedDict
 
 from backend.api.features.library import db as library_db
-from backend.data.graph import Graph, Link, Node, create_graph
+from backend.api.features.store import db as store_db
+from backend.data.graph import (
+    Graph,
+    Link,
+    Node,
+    create_graph,
+    get_graph,
+    get_graph_all_versions,
+)
+from backend.util.exceptions import DatabaseError, NotFoundError
 
 from .service import (
     decompose_goal_external,
@@ -16,6 +26,75 @@ from .service import (
 
 logger = logging.getLogger(__name__)
 
+# Block ID for AgentExecutorBlock - used to identify sub-agent nodes
+AGENT_EXECUTOR_BLOCK_ID = "e189baac-8c20-45a1-94a7-55177ea42565"
+
+
+class ExecutionSummary(TypedDict):
+    """Summary of a single execution for quality assessment."""
+
+    status: str
+    correctness_score: NotRequired[float]
+    activity_summary: NotRequired[str]
+
+
+class LibraryAgentSummary(TypedDict):
+    """Summary of a library agent for sub-agent composition.
+
+    Includes recent executions to help the LLM decide whether to use this agent.
+    Each execution shows status, correctness_score (0-1), and activity_summary.
+    """
+
+    graph_id: str
+    graph_version: int
+    name: str
+    description: str
+    input_schema: dict[str, Any]
+    output_schema: dict[str, Any]
+    recent_executions: NotRequired[list[ExecutionSummary]]
+
+
+class MarketplaceAgentSummary(TypedDict):
+    """Summary of a marketplace agent for sub-agent composition."""
+
+    name: str
+    description: str
+    sub_heading: str
+    creator: str
+    is_marketplace_agent: bool
+
+
+class DecompositionStep(TypedDict, total=False):
+    """A single step in decomposed instructions."""
+
+    description: str
+    action: str
+    block_name: str
+    tool: str
+    name: str
+
+
+class DecompositionResult(TypedDict, total=False):
+    """Result from decompose_goal - can be instructions, questions, or error."""
+
+    type: str
+    steps: list[DecompositionStep]
+    questions: list[dict[str, Any]]
+    error: str
+    error_type: str
+
+
+AgentSummary = LibraryAgentSummary | MarketplaceAgentSummary | dict[str, Any]
+
+
+def _to_dict_list(
+    agents: list[AgentSummary] | list[dict[str, Any]] | None,
+) -> list[dict[str, Any]] | None:
+    """Convert typed agent summaries to plain dicts for external service calls."""
+    if agents is None:
+        return None
+    return [dict(a) for a in agents]
+
 
 class AgentGeneratorNotConfiguredError(Exception):
     """Raised when the external Agent Generator service is not configured."""
@@ -36,15 +115,421 @@ def _check_service_configured() -> None:
         )
 
 
-async def decompose_goal(description: str, context: str = "") -> dict[str, Any] | None:
+_UUID_PATTERN = 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}",
+    re.IGNORECASE,
+)
+
+
+def extract_uuids_from_text(text: str) -> list[str]:
+    """Extract all UUID v4 strings from text.
+
+    Args:
+        text: Text that may contain UUIDs (e.g., user's goal description)
+
+    Returns:
+        List of unique UUIDs found in the text (lowercase)
+    """
+    matches = _UUID_PATTERN.findall(text)
+    return list({m.lower() for m in matches})
+
+
+async def get_library_agent_by_id(
+    user_id: str, agent_id: str
+) -> LibraryAgentSummary | None:
+    """Fetch a specific library agent by its ID (library agent ID or graph_id).
+
+    This function tries multiple lookup strategies:
+    1. First tries to find by graph_id (AgentGraph primary key)
+    2. If not found, tries to find by library agent ID (LibraryAgent primary key)
+
+    This handles both cases:
+    - User provides graph_id (e.g., from AgentExecutorBlock)
+    - User provides library agent ID (e.g., from library URL)
+
+    Args:
+        user_id: The user ID
+        agent_id: The ID to look up (can be graph_id or library agent ID)
+
+    Returns:
+        LibraryAgentSummary if found, None otherwise
+    """
+    try:
+        agent = await library_db.get_library_agent_by_graph_id(user_id, agent_id)
+        if agent:
+            logger.debug(f"Found library agent by graph_id: {agent.name}")
+            return LibraryAgentSummary(
+                graph_id=agent.graph_id,
+                graph_version=agent.graph_version,
+                name=agent.name,
+                description=agent.description,
+                input_schema=agent.input_schema,
+                output_schema=agent.output_schema,
+            )
+    except DatabaseError:
+        raise
+    except Exception as e:
+        logger.debug(f"Could not fetch library agent by graph_id {agent_id}: {e}")
+
+    try:
+        agent = await library_db.get_library_agent(agent_id, user_id)
+        if agent:
+            logger.debug(f"Found library agent by library_id: {agent.name}")
+            return LibraryAgentSummary(
+                graph_id=agent.graph_id,
+                graph_version=agent.graph_version,
+                name=agent.name,
+                description=agent.description,
+                input_schema=agent.input_schema,
+                output_schema=agent.output_schema,
+            )
+    except NotFoundError:
+        logger.debug(f"Library agent not found by library_id: {agent_id}")
+    except DatabaseError:
+        raise
+    except Exception as e:
+        logger.warning(
+            f"Could not fetch library agent by library_id {agent_id}: {e}",
+            exc_info=True,
+        )
+
+    return None
+
+
+get_library_agent_by_graph_id = get_library_agent_by_id
+
+
+async def get_library_agents_for_generation(
+    user_id: str,
+    search_query: str | None = None,
+    exclude_graph_id: str | None = None,
+    max_results: int = 15,
+) -> list[LibraryAgentSummary]:
+    """Fetch user's library agents formatted for Agent Generator.
+
+    Uses search-based fetching to return relevant agents instead of all agents.
+    This is more scalable for users with large libraries.
+
+    Includes recent_executions list to help the LLM assess agent quality:
+    - Each execution has status, correctness_score (0-1), and activity_summary
+    - This gives the LLM concrete examples of recent performance
+
+    Args:
+        user_id: The user ID
+        search_query: Optional search term to find relevant agents (user's goal/description)
+        exclude_graph_id: Optional graph ID to exclude (prevents circular references)
+        max_results: Maximum number of agents to return (default 15)
+
+    Returns:
+        List of LibraryAgentSummary with schemas and recent executions for sub-agent composition
+    """
+    try:
+        response = await library_db.list_library_agents(
+            user_id=user_id,
+            search_term=search_query,
+            page=1,
+            page_size=max_results,
+            include_executions=True,
+        )
+
+        results: list[LibraryAgentSummary] = []
+        for agent in response.agents:
+            if exclude_graph_id is not None and agent.graph_id == exclude_graph_id:
+                continue
+
+            summary = LibraryAgentSummary(
+                graph_id=agent.graph_id,
+                graph_version=agent.graph_version,
+                name=agent.name,
+                description=agent.description,
+                input_schema=agent.input_schema,
+                output_schema=agent.output_schema,
+            )
+            if agent.recent_executions:
+                exec_summaries: list[ExecutionSummary] = []
+                for ex in agent.recent_executions:
+                    exec_sum = ExecutionSummary(status=ex.status)
+                    if ex.correctness_score is not None:
+                        exec_sum["correctness_score"] = ex.correctness_score
+                    if ex.activity_summary:
+                        exec_sum["activity_summary"] = ex.activity_summary
+                    exec_summaries.append(exec_sum)
+                summary["recent_executions"] = exec_summaries
+            results.append(summary)
+        return results
+    except DatabaseError:
+        raise
+    except Exception as e:
+        logger.warning(f"Failed to fetch library agents: {e}")
+        return []
+
+
+async def search_marketplace_agents_for_generation(
+    search_query: str,
+    max_results: int = 10,
+) -> list[MarketplaceAgentSummary]:
+    """Search marketplace agents formatted for Agent Generator.
+
+    Note: This returns basic agent info. Full input/output schemas would require
+    additional graph fetches and is a potential future enhancement.
+
+    Args:
+        search_query: Search term to find relevant public agents
+        max_results: Maximum number of agents to return (default 10)
+
+    Returns:
+        List of MarketplaceAgentSummary (without detailed schemas for now)
+    """
+    try:
+        response = await store_db.get_store_agents(
+            search_query=search_query,
+            page=1,
+            page_size=max_results,
+        )
+
+        results: list[MarketplaceAgentSummary] = []
+        for agent in response.agents:
+            results.append(
+                MarketplaceAgentSummary(
+                    name=agent.agent_name,
+                    description=agent.description,
+                    sub_heading=agent.sub_heading,
+                    creator=agent.creator,
+                    is_marketplace_agent=True,
+                )
+            )
+        return results
+    except Exception as e:
+        logger.warning(f"Failed to search marketplace agents: {e}")
+        return []
+
+
+async def get_all_relevant_agents_for_generation(
+    user_id: str,
+    search_query: str | None = None,
+    exclude_graph_id: str | None = None,
+    include_library: bool = True,
+    include_marketplace: bool = True,
+    max_library_results: int = 15,
+    max_marketplace_results: int = 10,
+) -> list[AgentSummary]:
+    """Fetch relevant agents from library and/or marketplace.
+
+    Searches both user's library and marketplace by default.
+    Explicitly mentioned UUIDs in the search query are always looked up.
+
+    Args:
+        user_id: The user ID
+        search_query: Search term to find relevant agents (user's goal/description)
+        exclude_graph_id: Optional graph ID to exclude (prevents circular references)
+        include_library: Whether to search user's library (default True)
+        include_marketplace: Whether to also search marketplace (default True)
+        max_library_results: Max library agents to return (default 15)
+        max_marketplace_results: Max marketplace agents to return (default 10)
+
+    Returns:
+        List of AgentSummary, library agents first (with full schemas),
+        then marketplace agents (basic info only)
+    """
+    agents: list[AgentSummary] = []
+    seen_graph_ids: set[str] = set()
+
+    if search_query:
+        mentioned_uuids = extract_uuids_from_text(search_query)
+        for graph_id in mentioned_uuids:
+            if graph_id == exclude_graph_id:
+                continue
+            try:
+                agent = await get_library_agent_by_graph_id(user_id, graph_id)
+            except Exception as e:
+                logger.warning(
+                    f"Failed to fetch explicitly mentioned agent {graph_id}: {e}",
+                    exc_info=True,
+                )
+                continue
+            agent_graph_id = agent.get("graph_id") if agent else None
+            if agent and agent_graph_id and agent_graph_id not in seen_graph_ids:
+                agents.append(agent)
+                seen_graph_ids.add(agent_graph_id)
+                logger.debug(
+                    f"Found explicitly mentioned agent: {agent.get('name') or 'Unknown'}"
+                )
+
+    if include_library:
+        library_agents = await get_library_agents_for_generation(
+            user_id=user_id,
+            search_query=search_query,
+            exclude_graph_id=exclude_graph_id,
+            max_results=max_library_results,
+        )
+        for agent in library_agents:
+            graph_id = agent.get("graph_id")
+            if graph_id and graph_id not in seen_graph_ids:
+                agents.append(agent)
+                seen_graph_ids.add(graph_id)
+
+    if include_marketplace and search_query:
+        marketplace_agents = await search_marketplace_agents_for_generation(
+            search_query=search_query,
+            max_results=max_marketplace_results,
+        )
+        library_names: set[str] = set()
+        for a in agents:
+            name = a.get("name")
+            if name and isinstance(name, str):
+                library_names.add(name.lower())
+        for agent in marketplace_agents:
+            agent_name = agent.get("name")
+            if agent_name and isinstance(agent_name, str):
+                if agent_name.lower() not in library_names:
+                    agents.append(agent)
+
+    return agents
+
+
+def extract_search_terms_from_steps(
+    decomposition_result: DecompositionResult | dict[str, Any],
+) -> list[str]:
+    """Extract search terms from decomposed instruction steps.
+
+    Analyzes the decomposition result to extract relevant keywords
+    for additional library agent searches.
+
+    Args:
+        decomposition_result: Result from decompose_goal containing steps
+
+    Returns:
+        List of unique search terms extracted from steps
+    """
+    search_terms: list[str] = []
+
+    if decomposition_result.get("type") != "instructions":
+        return search_terms
+
+    steps = decomposition_result.get("steps", [])
+    if not steps:
+        return search_terms
+
+    step_keys: list[str] = ["description", "action", "block_name", "tool", "name"]
+
+    for step in steps:
+        for key in step_keys:
+            value = step.get(key)  # type: ignore[union-attr]
+            if isinstance(value, str) and len(value) > 3:
+                search_terms.append(value)
+
+    seen: set[str] = set()
+    unique_terms: list[str] = []
+    for term in search_terms:
+        term_lower = term.lower()
+        if term_lower not in seen:
+            seen.add(term_lower)
+            unique_terms.append(term)
+
+    return unique_terms
+
+
+async def enrich_library_agents_from_steps(
+    user_id: str,
+    decomposition_result: DecompositionResult | dict[str, Any],
+    existing_agents: list[AgentSummary] | list[dict[str, Any]],
+    exclude_graph_id: str | None = None,
+    include_marketplace: bool = True,
+    max_additional_results: int = 10,
+) -> list[AgentSummary] | list[dict[str, Any]]:
+    """Enrich library agents list with additional searches based on decomposed steps.
+
+    This implements two-phase search: after decomposition, we search for additional
+    relevant agents based on the specific steps identified.
+
+    Args:
+        user_id: The user ID
+        decomposition_result: Result from decompose_goal containing steps
+        existing_agents: Already fetched library agents from initial search
+        exclude_graph_id: Optional graph ID to exclude
+        include_marketplace: Whether to also search marketplace
+        max_additional_results: Max additional agents per search term (default 10)
+
+    Returns:
+        Combined list of library agents (existing + newly discovered)
+    """
+    search_terms = extract_search_terms_from_steps(decomposition_result)
+
+    if not search_terms:
+        return existing_agents
+
+    existing_ids: set[str] = set()
+    existing_names: set[str] = set()
+
+    for agent in existing_agents:
+        agent_name = agent.get("name")
+        if agent_name and isinstance(agent_name, str):
+            existing_names.add(agent_name.lower())
+        graph_id = agent.get("graph_id")  # type: ignore[call-overload]
+        if graph_id and isinstance(graph_id, str):
+            existing_ids.add(graph_id)
+
+    all_agents: list[AgentSummary] | list[dict[str, Any]] = list(existing_agents)
+
+    for term in search_terms[:3]:
+        try:
+            additional_agents = await get_all_relevant_agents_for_generation(
+                user_id=user_id,
+                search_query=term,
+                exclude_graph_id=exclude_graph_id,
+                include_marketplace=include_marketplace,
+                max_library_results=max_additional_results,
+                max_marketplace_results=5,
+            )
+
+            for agent in additional_agents:
+                agent_name = agent.get("name")
+                if not agent_name or not isinstance(agent_name, str):
+                    continue
+                agent_name_lower = agent_name.lower()
+
+                if agent_name_lower in existing_names:
+                    continue
+
+                graph_id = agent.get("graph_id")  # type: ignore[call-overload]
+                if graph_id and graph_id in existing_ids:
+                    continue
+
+                all_agents.append(agent)
+                existing_names.add(agent_name_lower)
+                if graph_id and isinstance(graph_id, str):
+                    existing_ids.add(graph_id)
+
+        except DatabaseError:
+            logger.error(f"Database error searching for agents with term '{term}'")
+            raise
+        except Exception as e:
+            logger.warning(
+                f"Failed to search for additional agents with term '{term}': {e}"
+            )
+
+    logger.debug(
+        f"Enriched library agents: {len(existing_agents)} initial + "
+        f"{len(all_agents) - len(existing_agents)} additional = {len(all_agents)} total"
+    )
+
+    return all_agents
+
+
+async def decompose_goal(
+    description: str,
+    context: str = "",
+    library_agents: list[AgentSummary] | None = None,
+) -> DecompositionResult | 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)
+        library_agents: User's library agents available for sub-agent composition
 
     Returns:
-        Dict with either:
+        DecompositionResult with either:
         - {"type": "clarifying_questions", "questions": [...]}
         - {"type": "instructions", "steps": [...]}
         Or None on error
@@ -54,14 +539,21 @@ async def decompose_goal(description: str, context: str = "") -> dict[str, Any]
     """
     _check_service_configured()
     logger.info("Calling external Agent Generator service for decompose_goal")
-    return await decompose_goal_external(description, context)
+    result = await decompose_goal_external(
+        description, context, _to_dict_list(library_agents)
+    )
+    return result  # type: ignore[return-value]
 
 
-async def generate_agent(instructions: dict[str, Any]) -> dict[str, Any] | None:
+async def generate_agent(
+    instructions: DecompositionResult | dict[str, Any],
+    library_agents: list[AgentSummary] | list[dict[str, Any]] | None = None,
+) -> dict[str, Any] | None:
     """Generate agent JSON from instructions.
 
     Args:
         instructions: Structured instructions from decompose_goal
+        library_agents: User's library agents available for sub-agent composition
 
     Returns:
         Agent JSON dict, error dict {"type": "error", ...}, or None on error
@@ -71,12 +563,12 @@ async def generate_agent(instructions: dict[str, Any]) -> dict[str, Any] | None:
     """
     _check_service_configured()
     logger.info("Calling external Agent Generator service for generate_agent")
-    result = await generate_agent_external(instructions)
+    result = await generate_agent_external(
+        dict(instructions), _to_dict_list(library_agents)
+    )
     if result:
-        # Check if it's an error response - pass through as-is
         if isinstance(result, dict) and result.get("type") == "error":
             return result
-        # Ensure required fields for successful agent generation
         if "id" not in result:
             result["id"] = str(uuid.uuid4())
         if "version" not in result:
@@ -86,6 +578,12 @@ async def generate_agent(instructions: dict[str, Any]) -> dict[str, Any] | None:
     return result
 
 
+class AgentJsonValidationError(Exception):
+    """Raised when agent JSON is invalid or missing required fields."""
+
+    pass
+
+
 def json_to_graph(agent_json: dict[str, Any]) -> Graph:
     """Convert agent JSON dict to Graph model.
 
@@ -94,25 +592,55 @@ def json_to_graph(agent_json: dict[str, Any]) -> Graph:
 
     Returns:
         Graph ready for saving
+
+    Raises:
+        AgentJsonValidationError: If required fields are missing from nodes or links
     """
     nodes = []
-    for n in agent_json.get("nodes", []):
+    for idx, n in enumerate(agent_json.get("nodes", [])):
+        block_id = n.get("block_id")
+        if not block_id:
+            node_id = n.get("id", f"index_{idx}")
+            raise AgentJsonValidationError(
+                f"Node '{node_id}' is missing required field 'block_id'"
+            )
         node = Node(
             id=n.get("id", str(uuid.uuid4())),
-            block_id=n["block_id"],
+            block_id=block_id,
             input_default=n.get("input_default", {}),
             metadata=n.get("metadata", {}),
         )
         nodes.append(node)
 
     links = []
-    for link_data in agent_json.get("links", []):
+    for idx, link_data in enumerate(agent_json.get("links", [])):
+        source_id = link_data.get("source_id")
+        sink_id = link_data.get("sink_id")
+        source_name = link_data.get("source_name")
+        sink_name = link_data.get("sink_name")
+
+        missing_fields = []
+        if not source_id:
+            missing_fields.append("source_id")
+        if not sink_id:
+            missing_fields.append("sink_id")
+        if not source_name:
+            missing_fields.append("source_name")
+        if not sink_name:
+            missing_fields.append("sink_name")
+
+        if missing_fields:
+            link_id = link_data.get("id", f"index_{idx}")
+            raise AgentJsonValidationError(
+                f"Link '{link_id}' is missing required fields: {', '.join(missing_fields)}"
+            )
+
         link = Link(
             id=link_data.get("id", str(uuid.uuid4())),
-            source_id=link_data["source_id"],
+            source_id=source_id,
-            sink_id=link_data["sink_id"],
+            sink_id=sink_id,
-            source_name=link_data["source_name"],
+            source_name=source_name,
-            sink_name=link_data["sink_name"],
+            sink_name=sink_name,
             is_static=link_data.get("is_static", False),
         )
         links.append(link)
@@ -133,22 +661,40 @@ def _reassign_node_ids(graph: Graph) -> None:
 
     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
+        link.id = str(uuid.uuid4())
         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]
 
 
+def _populate_agent_executor_user_ids(agent_json: dict[str, Any], user_id: str) -> None:
+    """Populate user_id in AgentExecutorBlock nodes.
+
+    The external agent generator creates AgentExecutorBlock nodes with empty user_id.
+    This function fills in the actual user_id so sub-agents run with correct permissions.
+
+    Args:
+        agent_json: Agent JSON dict (modified in place)
+        user_id: User ID to set
+    """
+    for node in agent_json.get("nodes", []):
+        if node.get("block_id") == AGENT_EXECUTOR_BLOCK_ID:
+            input_default = node.get("input_default", {})
+            if not input_default.get("user_id"):
+                input_default["user_id"] = user_id
+                node["input_default"] = input_default
+                logger.debug(
+                    f"Set user_id for AgentExecutorBlock node {node.get('id')}"
+                )
+
+
 async def save_agent_to_library(
     agent_json: dict[str, Any], user_id: str, is_update: bool = False
 ) -> tuple[Graph, Any]:
@@ -162,33 +708,27 @@ async def save_agent_to_library(
     Returns:
         Tuple of (created Graph, LibraryAgent)
     """
-    from backend.data.graph import get_graph_all_versions
+    # Populate user_id in AgentExecutorBlock nodes before conversion
+    _populate_agent_executor_user_ids(agent_json, user_id)
 
     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,
@@ -200,25 +740,31 @@ async def save_agent_to_library(
 
 
 async def get_agent_as_json(
-    graph_id: str, user_id: str | None
+    agent_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
+        agent_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
+    graph = await get_graph(agent_id, version=None, user_id=user_id)
+
+    if not graph and user_id:
+        try:
+            library_agent = await library_db.get_library_agent(agent_id, user_id)
+            graph = await get_graph(
+                library_agent.graph_id, version=None, user_id=user_id
+            )
+        except NotFoundError:
+            pass
 
-    # 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(
@@ -256,7 +802,9 @@ async def get_agent_as_json(
 
 
 async def generate_agent_patch(
-    update_request: str, current_agent: dict[str, Any]
+    update_request: str,
+    current_agent: dict[str, Any],
+    library_agents: list[AgentSummary] | None = None,
 ) -> dict[str, Any] | None:
     """Update an existing agent using natural language.
 
@@ -268,6 +816,7 @@ async def generate_agent_patch(
     Args:
         update_request: Natural language description of changes
         current_agent: Current agent JSON
+        library_agents: User's library agents available for sub-agent composition
 
     Returns:
         Updated agent JSON, clarifying questions dict {"type": "clarifying_questions", ...},
@@ -278,4 +827,6 @@ async def generate_agent_patch(
     """
     _check_service_configured()
     logger.info("Calling external Agent Generator service for generate_agent_patch")
-    return await generate_agent_patch_external(update_request, current_agent)
+    return await generate_agent_patch_external(
+        update_request, current_agent, _to_dict_list(library_agents)
+    )

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

@@ -1,11 +1,43 @@
 """Error handling utilities for agent generator."""
 
+import re
+
+
+def _sanitize_error_details(details: str) -> str:
+    """Sanitize error details to remove sensitive information.
+
+    Strips common patterns that could expose internal system info:
+    - File paths (Unix and Windows)
+    - Database connection strings
+    - URLs with credentials
+    - Stack trace internals
+
+    Args:
+        details: Raw error details string
+
+    Returns:
+        Sanitized error details safe for user display
+    """
+    sanitized = re.sub(
+        r"/[a-zA-Z0-9_./\-]+\.(py|js|ts|json|yaml|yml)", "[path]", details
+    )
+    sanitized = re.sub(r"[A-Z]:\\[a-zA-Z0-9_\\.\\-]+", "[path]", sanitized)
+    sanitized = re.sub(
+        r"(postgres|mysql|mongodb|redis)://[^\s]+", "[database_url]", sanitized
+    )
+    sanitized = re.sub(r"https?://[^:]+:[^@]+@[^\s]+", "[url]", sanitized)
+    sanitized = re.sub(r", line \d+", "", sanitized)
+    sanitized = re.sub(r'File "[^"]+",?', "", sanitized)
+
+    return sanitized.strip()
+
 
 def get_user_message_for_error(
     error_type: str,
     operation: str = "process the request",
     llm_parse_message: str | None = None,
     validation_message: str | None = None,
+    error_details: str | None = None,
 ) -> str:
     """Get a user-friendly error message based on error type.
 
@@ -19,25 +51,45 @@ def get_user_message_for_error(
             message (e.g., "analyze the goal", "generate the agent")
         llm_parse_message: Custom message for llm_parse_error type
         validation_message: Custom message for validation_error type
+        error_details: Optional additional details about the error
 
     Returns:
         User-friendly error message suitable for display to the user
     """
+    base_message = ""
+
     if error_type == "llm_parse_error":
-        return (
+        base_message = (
             llm_parse_message
             or "The AI had trouble processing this request. Please try again."
         )
     elif error_type == "validation_error":
-        return (
+        base_message = (
             validation_message
-            or "The request failed validation. Please try rephrasing."
+            or "The generated agent failed validation. "
+            "This usually happens when the agent structure doesn't match "
+            "what the platform expects. Please try simplifying your goal "
+            "or breaking it into smaller parts."
         )
     elif error_type == "patch_error":
-        return "Failed to apply the changes. Please try a different approach."
+        base_message = (
+            "Failed to apply the changes. The modification couldn't be "
+            "validated. Please try a different approach or simplify the change."
+        )
     elif error_type in ("timeout", "llm_timeout"):
-        return "The request took too long. Please try again."
+        base_message = (
+            "The request took too long to process. This can happen with "
+            "complex agents. Please try again or simplify your goal."
+        )
     elif error_type in ("rate_limit", "llm_rate_limit"):
-        return "The service is currently busy. Please try again in a moment."
+        base_message = "The service is currently busy. Please try again in a moment."
     else:
-        return f"Failed to {operation}. Please try again."
+        base_message = f"Failed to {operation}. Please try again."
+
+    if error_details:
+        details = _sanitize_error_details(error_details)
+        if len(details) > 200:
+            details = details[:200] + "..."
+        base_message += f"\n\nTechnical details: {details}"
+
+    return base_message

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

@@ -117,13 +117,16 @@ def _get_client() -> httpx.AsyncClient:
 
 
 async def decompose_goal_external(
-    description: str, context: str = ""
+    description: str,
+    context: str = "",
+    library_agents: list[dict[str, Any]] | None = None,
 ) -> dict[str, Any] | None:
     """Call the external service to decompose a goal.
 
     Args:
         description: Natural language goal description
         context: Additional context (e.g., answers to previous questions)
+        library_agents: User's library agents available for sub-agent composition
 
     Returns:
         Dict with either:
@@ -141,6 +144,8 @@ async def decompose_goal_external(
     if context:
         # The external service uses user_instruction for additional context
         payload["user_instruction"] = context
+    if library_agents:
+        payload["library_agents"] = library_agents
 
     try:
         response = await client.post("/api/decompose-description", json=payload)
@@ -207,21 +212,25 @@ async def decompose_goal_external(
 
 async def generate_agent_external(
     instructions: dict[str, Any],
+    library_agents: list[dict[str, Any]] | None = None,
 ) -> dict[str, Any] | None:
     """Call the external service to generate an agent from instructions.
 
     Args:
         instructions: Structured instructions from decompose_goal
+        library_agents: User's library agents available for sub-agent composition
 
     Returns:
         Agent JSON dict on success, or error dict {"type": "error", ...} on error
     """
     client = _get_client()
 
+    payload: dict[str, Any] = {"instructions": instructions}
+    if library_agents:
+        payload["library_agents"] = library_agents
+
     try:
-        response = await client.post(
+        response = await client.post("/api/generate-agent", json=payload)
-            "/api/generate-agent", json={"instructions": instructions}
-        )
         response.raise_for_status()
         data = response.json()
 
@@ -229,8 +238,7 @@ async def generate_agent_external(
             error_msg = data.get("error", "Unknown error from Agent Generator")
             error_type = data.get("error_type", "unknown")
             logger.error(
-                f"Agent Generator generation failed: {error_msg} "
+                f"Agent Generator generation failed: {error_msg} (type: {error_type})"
-                f"(type: {error_type})"
             )
             return _create_error_response(error_msg, error_type)
 
@@ -251,27 +259,31 @@ async def generate_agent_external(
 
 
 async def generate_agent_patch_external(
-    update_request: str, current_agent: dict[str, Any]
+    update_request: str,
+    current_agent: dict[str, Any],
+    library_agents: list[dict[str, Any]] | None = None,
 ) -> dict[str, Any] | None:
     """Call the external service to generate a patch for an existing agent.
 
     Args:
         update_request: Natural language description of changes
         current_agent: Current agent JSON
+        library_agents: User's library agents available for sub-agent composition
 
     Returns:
         Updated agent JSON, clarifying questions dict, or error dict on error
     """
     client = _get_client()
 
+    payload: dict[str, Any] = {
+        "update_request": update_request,
+        "current_agent_json": current_agent,
+    }
+    if library_agents:
+        payload["library_agents"] = library_agents
+
     try:
-        response = await client.post(
+        response = await client.post("/api/update-agent", json=payload)
-            "/api/update-agent",
-            json={
-                "update_request": update_request,
-                "current_agent_json": current_agent,
-            },
-        )
         response.raise_for_status()
         data = response.json()
 

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

@@ -1,6 +1,7 @@
 """Shared agent search functionality for find_agent and find_library_agent tools."""
 
 import logging
+import re
 from typing import Literal
 
 from backend.api.features.library import db as library_db
@@ -19,6 +20,85 @@ logger = logging.getLogger(__name__)
 
 SearchSource = Literal["marketplace", "library"]
 
+_UUID_PATTERN = 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}$",
+    re.IGNORECASE,
+)
+
+
+def _is_uuid(text: str) -> bool:
+    """Check if text is a valid UUID v4."""
+    return bool(_UUID_PATTERN.match(text.strip()))
+
+
+async def _get_library_agent_by_id(user_id: str, agent_id: str) -> AgentInfo | None:
+    """Fetch a library agent by ID (library agent ID or graph_id).
+
+    Tries multiple lookup strategies:
+    1. First by graph_id (AgentGraph primary key)
+    2. Then by library agent ID (LibraryAgent primary key)
+
+    Args:
+        user_id: The user ID
+        agent_id: The ID to look up (can be graph_id or library agent ID)
+
+    Returns:
+        AgentInfo if found, None otherwise
+    """
+    try:
+        agent = await library_db.get_library_agent_by_graph_id(user_id, agent_id)
+        if agent:
+            logger.debug(f"Found library agent by graph_id: {agent.name}")
+            return 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,
+            )
+    except DatabaseError:
+        raise
+    except Exception as e:
+        logger.warning(
+            f"Could not fetch library agent by graph_id {agent_id}: {e}",
+            exc_info=True,
+        )
+
+    try:
+        agent = await library_db.get_library_agent(agent_id, user_id)
+        if agent:
+            logger.debug(f"Found library agent by library_id: {agent.name}")
+            return 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,
+            )
+    except NotFoundError:
+        logger.debug(f"Library agent not found by library_id: {agent_id}")
+    except DatabaseError:
+        raise
+    except Exception as e:
+        logger.warning(
+            f"Could not fetch library agent by library_id {agent_id}: {e}",
+            exc_info=True,
+        )
+
+    return None
+
 
 async def search_agents(
     query: str,
@@ -69,29 +149,37 @@ async def search_agents(
                         is_featured=False,
                     )
                 )
-        else:  # library
+        else:
-            logger.info(f"Searching user library for: {query}")
+            if _is_uuid(query):
-            results = await library_db.list_library_agents(
+                logger.info(f"Query looks like UUID, trying direct lookup: {query}")
-                user_id=user_id,  # type: ignore[arg-type]
+                agent = await _get_library_agent_by_id(user_id, query)  # type: ignore[arg-type]
-                search_term=query,
+                if agent:
-                page_size=10,
+                    agents.append(agent)
-            )
+                    logger.info(f"Found agent by direct ID lookup: {agent.name}")
-            for agent in results.agents:
+
-                agents.append(
+            if not agents:
-                    AgentInfo(
+                logger.info(f"Searching user library for: {query}")
-                        id=agent.id,
+                results = await library_db.list_library_agents(
-                        name=agent.name,
+                    user_id=user_id,  # type: ignore[arg-type]
-                        description=agent.description or "",
+                    search_term=query,
-                        source="library",
+                    page_size=10,
-                        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,
-                    )
                 )
+                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

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

@@ -8,7 +8,9 @@ from backend.api.features.chat.model import ChatSession
 from .agent_generator import (
     AgentGeneratorNotConfiguredError,
     decompose_goal,
+    enrich_library_agents_from_steps,
     generate_agent,
+    get_all_relevant_agents_for_generation,
     get_user_message_for_error,
     save_agent_to_library,
 )
@@ -103,9 +105,24 @@ class CreateAgentTool(BaseTool):
                 session_id=session_id,
             )
 
-        # Step 1: Decompose goal into steps
+        library_agents = None
+        if user_id:
+            try:
+                library_agents = await get_all_relevant_agents_for_generation(
+                    user_id=user_id,
+                    search_query=description,
+                    include_marketplace=True,
+                )
+                logger.debug(
+                    f"Found {len(library_agents)} relevant agents for sub-agent composition"
+                )
+            except Exception as e:
+                logger.warning(f"Failed to fetch library agents: {e}")
+
         try:
-            decomposition_result = await decompose_goal(description, context)
+            decomposition_result = await decompose_goal(
+                description, context, library_agents
+            )
         except AgentGeneratorNotConfiguredError:
             return ErrorResponse(
                 message=(
@@ -124,7 +141,6 @@ class CreateAgentTool(BaseTool):
                 session_id=session_id,
             )
 
-        # Check if the result is an error from the external service
         if decomposition_result.get("type") == "error":
             error_msg = decomposition_result.get("error", "Unknown error")
             error_type = decomposition_result.get("error_type", "unknown")
@@ -144,7 +160,6 @@ class CreateAgentTool(BaseTool):
                 session_id=session_id,
             )
 
-        # Check if LLM returned clarifying questions
         if decomposition_result.get("type") == "clarifying_questions":
             questions = decomposition_result.get("questions", [])
             return ClarificationNeededResponse(
@@ -163,7 +178,6 @@ class CreateAgentTool(BaseTool):
                 session_id=session_id,
             )
 
-        # Check for unachievable/vague goals
         if decomposition_result.get("type") == "unachievable_goal":
             suggested = decomposition_result.get("suggested_goal", "")
             reason = decomposition_result.get("reason", "")
@@ -190,9 +204,22 @@ class CreateAgentTool(BaseTool):
                 session_id=session_id,
             )
 
-        # Step 2: Generate agent JSON (external service handles fixing and validation)
+        if user_id and library_agents is not None:
+            try:
+                library_agents = await enrich_library_agents_from_steps(
+                    user_id=user_id,
+                    decomposition_result=decomposition_result,
+                    existing_agents=library_agents,
+                    include_marketplace=True,
+                )
+                logger.debug(
+                    f"After enrichment: {len(library_agents)} total agents for sub-agent composition"
+                )
+            except Exception as e:
+                logger.warning(f"Failed to enrich library agents from steps: {e}")
+
         try:
-            agent_json = await generate_agent(decomposition_result)
+            agent_json = await generate_agent(decomposition_result, library_agents)
         except AgentGeneratorNotConfiguredError:
             return ErrorResponse(
                 message=(
@@ -211,7 +238,6 @@ class CreateAgentTool(BaseTool):
                 session_id=session_id,
             )
 
-        # Check if the result is an error from the external service
         if isinstance(agent_json, dict) and agent_json.get("type") == "error":
             error_msg = agent_json.get("error", "Unknown error")
             error_type = agent_json.get("error_type", "unknown")
@@ -219,7 +245,12 @@ class CreateAgentTool(BaseTool):
                 error_type,
                 operation="generate the agent",
                 llm_parse_message="The AI had trouble generating the agent. Please try again or simplify your goal.",
-                validation_message="The generated agent failed validation. Please try rephrasing your goal.",
+                validation_message=(
+                    "I wasn't able to create a valid agent for this request. "
+                    "The generated workflow had some structural issues. "
+                    "Please try simplifying your goal or breaking it into smaller steps."
+                ),
+                error_details=error_msg,
             )
             return ErrorResponse(
                 message=user_message,
@@ -237,7 +268,6 @@ class CreateAgentTool(BaseTool):
         node_count = len(agent_json.get("nodes", []))
         link_count = len(agent_json.get("links", []))
 
-        # Step 3: Preview or save
         if not save:
             return AgentPreviewResponse(
                 message=(
@@ -252,7 +282,6 @@ class CreateAgentTool(BaseTool):
                 session_id=session_id,
             )
 
-        # Save to library
         if not user_id:
             return ErrorResponse(
                 message="You must be logged in to save agents.",
@@ -270,7 +299,7 @@ class CreateAgentTool(BaseTool):
                 agent_id=created_graph.id,
                 agent_name=created_graph.name,
                 library_agent_id=library_agent.id,
-                library_agent_link=f"/library/{library_agent.id}",
+                library_agent_link=f"/library/agents/{library_agent.id}",
                 agent_page_link=f"/build?flowID={created_graph.id}",
                 session_id=session_id,
             )

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

@@ -9,6 +9,7 @@ from .agent_generator import (
     AgentGeneratorNotConfiguredError,
     generate_agent_patch,
     get_agent_as_json,
+    get_all_relevant_agents_for_generation,
     get_user_message_for_error,
     save_agent_to_library,
 )
@@ -117,7 +118,6 @@ class EditAgentTool(BaseTool):
                 session_id=session_id,
             )
 
-        # Step 1: Fetch current agent
         current_agent = await get_agent_as_json(agent_id, user_id)
 
         if current_agent is None:
@@ -127,14 +127,30 @@ class EditAgentTool(BaseTool):
                 session_id=session_id,
             )
 
-        # Build the update request with context
+        library_agents = None
+        if user_id:
+            try:
+                graph_id = current_agent.get("id")
+                library_agents = await get_all_relevant_agents_for_generation(
+                    user_id=user_id,
+                    search_query=changes,
+                    exclude_graph_id=graph_id,
+                    include_marketplace=True,
+                )
+                logger.debug(
+                    f"Found {len(library_agents)} relevant agents for sub-agent composition"
+                )
+            except Exception as e:
+                logger.warning(f"Failed to fetch library agents: {e}")
+
         update_request = changes
         if context:
             update_request = f"{changes}\n\nAdditional context:\n{context}"
 
-        # Step 2: Generate updated agent (external service handles fixing and validation)
         try:
-            result = await generate_agent_patch(update_request, current_agent)
+            result = await generate_agent_patch(
+                update_request, current_agent, library_agents
+            )
         except AgentGeneratorNotConfiguredError:
             return ErrorResponse(
                 message=(
@@ -153,7 +169,6 @@ class EditAgentTool(BaseTool):
                 session_id=session_id,
             )
 
-        # Check if the result is an error from the external service
         if isinstance(result, dict) and result.get("type") == "error":
             error_msg = result.get("error", "Unknown error")
             error_type = result.get("error_type", "unknown")
@@ -162,6 +177,7 @@ class EditAgentTool(BaseTool):
                 operation="generate the changes",
                 llm_parse_message="The AI had trouble generating the changes. Please try again or simplify your request.",
                 validation_message="The generated changes failed validation. Please try rephrasing your request.",
+                error_details=error_msg,
             )
             return ErrorResponse(
                 message=user_message,
@@ -175,7 +191,6 @@ class EditAgentTool(BaseTool):
                 session_id=session_id,
             )
 
-        # Check if LLM returned clarifying questions
         if result.get("type") == "clarifying_questions":
             questions = result.get("questions", [])
             return ClarificationNeededResponse(
@@ -194,7 +209,6 @@ class EditAgentTool(BaseTool):
                 session_id=session_id,
             )
 
-        # Result is the updated agent JSON
         updated_agent = result
 
         agent_name = updated_agent.get("name", "Updated Agent")
@@ -202,7 +216,6 @@ class EditAgentTool(BaseTool):
         node_count = len(updated_agent.get("nodes", []))
         link_count = len(updated_agent.get("links", []))
 
-        # Step 3: Preview or save
         if not save:
             return AgentPreviewResponse(
                 message=(
@@ -218,7 +231,6 @@ class EditAgentTool(BaseTool):
                 session_id=session_id,
             )
 
-        # Save to library (creates a new version)
         if not user_id:
             return ErrorResponse(
                 message="You must be logged in to save agents.",
@@ -236,7 +248,7 @@ class EditAgentTool(BaseTool):
                 agent_id=created_graph.id,
                 agent_name=created_graph.name,
                 library_agent_id=library_agent.id,
-                library_agent_link=f"/library/{library_agent.id}",
+                library_agent_link=f"/library/agents/{library_agent.id}",
                 agent_page_link=f"/build?flowID={created_graph.id}",
                 session_id=session_id,
             )

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

@@ -8,7 +8,7 @@ from backend.api.features.library import model as library_model
 from backend.api.features.store import db as store_db
 from backend.data import graph as graph_db
 from backend.data.graph import GraphModel
-from backend.data.model import CredentialsFieldInfo, CredentialsMetaInput
+from backend.data.model import Credentials, CredentialsFieldInfo, CredentialsMetaInput
 from backend.integrations.creds_manager import IntegrationCredentialsManager
 from backend.util.exceptions import NotFoundError
 
@@ -266,13 +266,14 @@ async def match_user_credentials_to_graph(
         credential_requirements,
         _node_fields,
     ) in aggregated_creds.items():
-        # Find first matching credential by provider and type
+        # Find first matching credential by provider, type, and scopes
         matching_cred = next(
             (
                 cred
                 for cred in available_creds
                 if cred.provider in credential_requirements.provider
                 and cred.type in credential_requirements.supported_types
+                and _credential_has_required_scopes(cred, credential_requirements)
             ),
             None,
         )
@@ -296,10 +297,17 @@ async def match_user_credentials_to_graph(
                     f"{credential_field_name} (validation failed: {e})"
                 )
         else:
+            # Build a helpful error message including scope requirements
+            error_parts = [
+                f"provider in {list(credential_requirements.provider)}",
+                f"type in {list(credential_requirements.supported_types)}",
+            ]
+            if credential_requirements.required_scopes:
+                error_parts.append(
+                    f"scopes including {list(credential_requirements.required_scopes)}"
+                )
             missing_creds.append(
-                f"{credential_field_name} "
+                f"{credential_field_name} (requires {', '.join(error_parts)})"
-                f"(requires provider in {list(credential_requirements.provider)}, "
-                f"type in {list(credential_requirements.supported_types)})"
             )
 
     logger.info(
@@ -309,6 +317,28 @@ async def match_user_credentials_to_graph(
     return graph_credentials_inputs, missing_creds
 
 
+def _credential_has_required_scopes(
+    credential: Credentials,
+    requirements: CredentialsFieldInfo,
+) -> bool:
+    """
+    Check if a credential has all the scopes required by the block.
+
+    For OAuth2 credentials, verifies that the credential's scopes are a superset
+    of the required scopes. For other credential types, returns True (no scope check).
+    """
+    # Only OAuth2 credentials have scopes to check
+    if credential.type != "oauth2":
+        return True
+
+    # If no scopes are required, any credential matches
+    if not requirements.required_scopes:
+        return True
+
+    # Check that credential scopes are a superset of required scopes
+    return set(credential.scopes).issuperset(requirements.required_scopes)
+
+
 async def check_user_has_required_credentials(
     user_id: str,
     required_credentials: list[CredentialsMetaInput],

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

@@ -39,6 +39,7 @@ 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 +50,9 @@ 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.
@@ -76,24 +80,32 @@ async def list_library_agents(
         "isArchived": False,
     }
 
-    # Build search filter if applicable
     if search_term:
-        where_clause["OR"] = [
+        words = [w.strip() for w in search_term.split() if len(w.strip()) >= 3]
-            {
+        if words:
-                "AgentGraph": {
+            or_conditions: list[prisma.types.LibraryAgentWhereInput] = []
-                    "is": {"name": {"contains": search_term, "mode": "insensitive"}}
+            for word in words:
-                }
+                or_conditions.append(
-            },
+                    {
-            {
+                        "AgentGraph": {
-                "AgentGraph": {
+                            "is": {"name": {"contains": word, "mode": "insensitive"}}
-                    "is": {
+                        }
-                        "description": {"contains": search_term, "mode": "insensitive"}
                     }
-                }
+                )
-            },
+                or_conditions.append(
-        ]
+                    {
+                        "AgentGraph": {
+                            "is": {
+                                "description": {
+                                    "contains": word,
+                                    "mode": "insensitive",
+                                }
+                            }
+                        }
+                    }
+                )
+            where_clause["OR"] = or_conditions
 
-    # Determine sorting
     order_by: prisma.types.LibraryAgentOrderByInput | None = None
 
     if sort_by == library_model.LibraryAgentSort.CREATED_AT:
@@ -105,7 +117,7 @@ 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,

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
+    COMPLETED = "COMPLETED"
-    HEALTHY = "HEALTHY"  # Agent is running (not all runs have completed)
+    HEALTHY = "HEALTHY"
-    WAITING = "WAITING"  # Agent is queued or waiting to start
+    WAITING = "WAITING"
-    ERROR = "ERROR"  # Agent is in an error state
+    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
-
+    execution_count: int = 0
-    # Whether the user can access the underlying graph
+    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
+        execution_count = len(executions)
-        can_access_graph = agent.AgentGraph.userId == agent.userId
+        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/blocks/llm.py

@@ -115,7 +115,6 @@ 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_3_HAIKU = "claude-3-haiku-20240307"
     # AI/ML API models
     AIML_API_QWEN2_5_72B = "Qwen/Qwen2.5-72B-Instruct-Turbo"
@@ -280,9 +279,6 @@ 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

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()
@@ -230,7 +230,7 @@ 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()
@@ -330,7 +330,7 @@ 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()

autogpt_platform/backend/backend/data/block_cost_config.py

@@ -81,7 +81,6 @@ MODEL_COST: dict[LlmModel, int] = {
     LlmModel.CLAUDE_4_5_HAIKU: 4,
     LlmModel.CLAUDE_4_5_OPUS: 14,
     LlmModel.CLAUDE_4_5_SONNET: 9,
-    LlmModel.CLAUDE_3_7_SONNET: 5,
     LlmModel.CLAUDE_3_HAIKU: 1,
     LlmModel.AIML_API_QWEN2_5_72B: 1,
     LlmModel.AIML_API_LLAMA3_1_70B: 1,

autogpt_platform/backend/backend/data/model.py

@@ -666,10 +666,16 @@ class CredentialsFieldInfo(BaseModel, Generic[CP, CT]):
         if not (self.discriminator and self.discriminator_mapping):
             return self
 
+        try:
+            provider = self.discriminator_mapping[discriminator_value]
+        except KeyError:
+            raise ValueError(
+                f"Model '{discriminator_value}' is not supported. "
+                "It may have been deprecated. Please update your agent configuration."
+            )
+
         return CredentialsFieldInfo(
-            credentials_provider=frozenset(
+            credentials_provider=frozenset([provider]),
-                [self.discriminator_mapping[discriminator_value]]
-            ),
             credentials_types=self.supported_types,
             credentials_scopes=self.required_scopes,
             discriminator=self.discriminator,

autogpt_platform/backend/migrations/20260126120000_migrate_claude_3_7_to_4_5_sonnet/migration.sql

@@ -0,0 +1,22 @@
+-- Migrate Claude 3.7 Sonnet to Claude 4.5 Sonnet
+-- This updates all AgentNode blocks that use the deprecated Claude 3.7 Sonnet model
+-- Anthropic is retiring claude-3-7-sonnet-20250219 on February 19, 2026
+
+-- Update AgentNode constant inputs
+UPDATE "AgentNode"
+SET    "constantInput" = JSONB_SET(
+         "constantInput"::jsonb,
+         '{model}',
+         '"claude-sonnet-4-5-20250929"'::jsonb
+       )
+WHERE  "constantInput"::jsonb->>'model' = 'claude-3-7-sonnet-20250219';
+
+-- Update AgentPreset input overrides (stored in AgentNodeExecutionInputOutput)
+UPDATE "AgentNodeExecutionInputOutput"
+SET    "data" = JSONB_SET(
+         "data"::jsonb,
+         '{model}',
+         '"claude-sonnet-4-5-20250929"'::jsonb
+       )
+WHERE  "agentPresetId" IS NOT NULL
+  AND  "data"::jsonb->>'model' = 'claude-3-7-sonnet-20250219';

autogpt_platform/backend/snapshots/lib_agts_search

@@ -31,6 +31,10 @@
       "has_sensitive_action": false,
       "trigger_setup_info": null,
       "new_output": false,
+      "execution_count": 0,
+      "success_rate": null,
+      "avg_correctness_score": null,
+      "recent_executions": [],
       "can_access_graph": true,
       "is_latest_version": true,
       "is_favorite": false,
@@ -72,6 +76,10 @@
       "has_sensitive_action": false,
       "trigger_setup_info": null,
       "new_output": false,
+      "execution_count": 0,
+      "success_rate": null,
+      "avg_correctness_score": null,
+      "recent_executions": [],
       "can_access_graph": false,
       "is_latest_version": true,
       "is_favorite": false,

autogpt_platform/backend/test/agent_generator/test_core_integration.py

@@ -57,7 +57,8 @@ class TestDecomposeGoal:
 
             result = await core.decompose_goal("Build a chatbot")
 
-            mock_external.assert_called_once_with("Build a chatbot", "")
+            # library_agents defaults to None
+            mock_external.assert_called_once_with("Build a chatbot", "", None)
             assert result == expected_result
 
     @pytest.mark.asyncio
@@ -74,7 +75,8 @@ class TestDecomposeGoal:
 
             await core.decompose_goal("Build a chatbot", "Use Python")
 
-            mock_external.assert_called_once_with("Build a chatbot", "Use Python")
+            # library_agents defaults to None
+            mock_external.assert_called_once_with("Build a chatbot", "Use Python", None)
 
     @pytest.mark.asyncio
     async def test_returns_none_on_service_failure(self):
@@ -109,7 +111,8 @@ class TestGenerateAgent:
             instructions = {"type": "instructions", "steps": ["Step 1"]}
             result = await core.generate_agent(instructions)
 
-            mock_external.assert_called_once_with(instructions)
+            # library_agents defaults to None
+            mock_external.assert_called_once_with(instructions, None)
             # Result should have id, version, is_active added if not present
             assert result is not None
             assert result["name"] == "Test Agent"
@@ -174,7 +177,8 @@ class TestGenerateAgentPatch:
             current_agent = {"nodes": [], "links": []}
             result = await core.generate_agent_patch("Add a node", current_agent)
 
-            mock_external.assert_called_once_with("Add a node", current_agent)
+            # library_agents defaults to None
+            mock_external.assert_called_once_with("Add a node", current_agent, None)
             assert result == expected_result
 
     @pytest.mark.asyncio

autogpt_platform/backend/test/agent_generator/test_library_agents.py

@@ -0,0 +1,841 @@
+"""
+Tests for library agent fetching functionality in agent generator.
+
+This test suite verifies the search-based library agent fetching,
+including the combination of library and marketplace agents.
+"""
+
+from unittest.mock import AsyncMock, MagicMock, patch
+
+import pytest
+
+from backend.api.features.chat.tools.agent_generator import core
+
+
+class TestGetLibraryAgentsForGeneration:
+    """Test get_library_agents_for_generation function."""
+
+    @pytest.mark.asyncio
+    async def test_fetches_agents_with_search_term(self):
+        """Test that search_term is passed to the library db."""
+        # Create a mock agent with proper attribute values
+        mock_agent = MagicMock()
+        mock_agent.graph_id = "agent-123"
+        mock_agent.graph_version = 1
+        mock_agent.name = "Email Agent"
+        mock_agent.description = "Sends emails"
+        mock_agent.input_schema = {"properties": {}}
+        mock_agent.output_schema = {"properties": {}}
+        mock_agent.recent_executions = []
+
+        mock_response = MagicMock()
+        mock_response.agents = [mock_agent]
+
+        with patch.object(
+            core.library_db,
+            "list_library_agents",
+            new_callable=AsyncMock,
+            return_value=mock_response,
+        ) as mock_list:
+            result = await core.get_library_agents_for_generation(
+                user_id="user-123",
+                search_query="send email",
+            )
+
+            mock_list.assert_called_once_with(
+                user_id="user-123",
+                search_term="send email",
+                page=1,
+                page_size=15,
+                include_executions=True,
+            )
+
+        # Verify result format
+        assert len(result) == 1
+        assert result[0]["graph_id"] == "agent-123"
+        assert result[0]["name"] == "Email Agent"
+
+    @pytest.mark.asyncio
+    async def test_excludes_specified_graph_id(self):
+        """Test that agents with excluded graph_id are filtered out."""
+        mock_response = MagicMock()
+        mock_response.agents = [
+            MagicMock(
+                graph_id="agent-123",
+                graph_version=1,
+                name="Agent 1",
+                description="First agent",
+                input_schema={},
+                output_schema={},
+                recent_executions=[],
+            ),
+            MagicMock(
+                graph_id="agent-456",
+                graph_version=1,
+                name="Agent 2",
+                description="Second agent",
+                input_schema={},
+                output_schema={},
+                recent_executions=[],
+            ),
+        ]
+
+        with patch.object(
+            core.library_db,
+            "list_library_agents",
+            new_callable=AsyncMock,
+            return_value=mock_response,
+        ):
+            result = await core.get_library_agents_for_generation(
+                user_id="user-123",
+                exclude_graph_id="agent-123",
+            )
+
+        # Verify the excluded agent is not in results
+        assert len(result) == 1
+        assert result[0]["graph_id"] == "agent-456"
+
+    @pytest.mark.asyncio
+    async def test_respects_max_results(self):
+        """Test that max_results parameter limits the page_size."""
+        mock_response = MagicMock()
+        mock_response.agents = []
+
+        with patch.object(
+            core.library_db,
+            "list_library_agents",
+            new_callable=AsyncMock,
+            return_value=mock_response,
+        ) as mock_list:
+            await core.get_library_agents_for_generation(
+                user_id="user-123",
+                max_results=5,
+            )
+
+            mock_list.assert_called_once_with(
+                user_id="user-123",
+                search_term=None,
+                page=1,
+                page_size=5,
+                include_executions=True,
+            )
+
+
+class TestSearchMarketplaceAgentsForGeneration:
+    """Test search_marketplace_agents_for_generation function."""
+
+    @pytest.mark.asyncio
+    async def test_searches_marketplace_with_query(self):
+        """Test that marketplace is searched with the query."""
+        mock_response = MagicMock()
+        mock_response.agents = [
+            MagicMock(
+                agent_name="Public Agent",
+                description="A public agent",
+                sub_heading="Does something useful",
+                creator="creator-1",
+            )
+        ]
+
+        # The store_db is dynamically imported, so patch the import path
+        with patch(
+            "backend.api.features.store.db.get_store_agents",
+            new_callable=AsyncMock,
+            return_value=mock_response,
+        ) as mock_search:
+            result = await core.search_marketplace_agents_for_generation(
+                search_query="automation",
+                max_results=10,
+            )
+
+            mock_search.assert_called_once_with(
+                search_query="automation",
+                page=1,
+                page_size=10,
+            )
+
+        assert len(result) == 1
+        assert result[0]["name"] == "Public Agent"
+        assert result[0]["is_marketplace_agent"] is True
+
+    @pytest.mark.asyncio
+    async def test_handles_marketplace_error_gracefully(self):
+        """Test that marketplace errors don't crash the function."""
+        with patch(
+            "backend.api.features.store.db.get_store_agents",
+            new_callable=AsyncMock,
+            side_effect=Exception("Marketplace unavailable"),
+        ):
+            result = await core.search_marketplace_agents_for_generation(
+                search_query="test"
+            )
+
+        # Should return empty list, not raise exception
+        assert result == []
+
+
+class TestGetAllRelevantAgentsForGeneration:
+    """Test get_all_relevant_agents_for_generation function."""
+
+    @pytest.mark.asyncio
+    async def test_combines_library_and_marketplace_agents(self):
+        """Test that agents from both sources are combined."""
+        library_agents = [
+            {
+                "graph_id": "lib-123",
+                "graph_version": 1,
+                "name": "Library Agent",
+                "description": "From library",
+                "input_schema": {},
+                "output_schema": {},
+            }
+        ]
+
+        marketplace_agents = [
+            {
+                "name": "Market Agent",
+                "description": "From marketplace",
+                "sub_heading": "Sub heading",
+                "creator": "creator-1",
+                "is_marketplace_agent": True,
+            }
+        ]
+
+        with patch.object(
+            core,
+            "get_library_agents_for_generation",
+            new_callable=AsyncMock,
+            return_value=library_agents,
+        ):
+            with patch.object(
+                core,
+                "search_marketplace_agents_for_generation",
+                new_callable=AsyncMock,
+                return_value=marketplace_agents,
+            ):
+                result = await core.get_all_relevant_agents_for_generation(
+                    user_id="user-123",
+                    search_query="test query",
+                    include_marketplace=True,
+                )
+
+        # Library agents should come first
+        assert len(result) == 2
+        assert result[0]["name"] == "Library Agent"
+        assert result[1]["name"] == "Market Agent"
+
+    @pytest.mark.asyncio
+    async def test_deduplicates_by_name(self):
+        """Test that marketplace agents with same name as library are excluded."""
+        library_agents = [
+            {
+                "graph_id": "lib-123",
+                "graph_version": 1,
+                "name": "Shared Agent",
+                "description": "From library",
+                "input_schema": {},
+                "output_schema": {},
+            }
+        ]
+
+        marketplace_agents = [
+            {
+                "name": "Shared Agent",  # Same name, should be deduplicated
+                "description": "From marketplace",
+                "sub_heading": "Sub heading",
+                "creator": "creator-1",
+                "is_marketplace_agent": True,
+            },
+            {
+                "name": "Unique Agent",
+                "description": "Only in marketplace",
+                "sub_heading": "Sub heading",
+                "creator": "creator-2",
+                "is_marketplace_agent": True,
+            },
+        ]
+
+        with patch.object(
+            core,
+            "get_library_agents_for_generation",
+            new_callable=AsyncMock,
+            return_value=library_agents,
+        ):
+            with patch.object(
+                core,
+                "search_marketplace_agents_for_generation",
+                new_callable=AsyncMock,
+                return_value=marketplace_agents,
+            ):
+                result = await core.get_all_relevant_agents_for_generation(
+                    user_id="user-123",
+                    search_query="test",
+                    include_marketplace=True,
+                )
+
+        # Shared Agent from marketplace should be excluded
+        assert len(result) == 2
+        names = [a["name"] for a in result]
+        assert "Shared Agent" in names
+        assert "Unique Agent" in names
+
+    @pytest.mark.asyncio
+    async def test_skips_marketplace_when_disabled(self):
+        """Test that marketplace is not searched when include_marketplace=False."""
+        library_agents = [
+            {
+                "graph_id": "lib-123",
+                "graph_version": 1,
+                "name": "Library Agent",
+                "description": "From library",
+                "input_schema": {},
+                "output_schema": {},
+            }
+        ]
+
+        with patch.object(
+            core,
+            "get_library_agents_for_generation",
+            new_callable=AsyncMock,
+            return_value=library_agents,
+        ):
+            with patch.object(
+                core,
+                "search_marketplace_agents_for_generation",
+                new_callable=AsyncMock,
+            ) as mock_marketplace:
+                result = await core.get_all_relevant_agents_for_generation(
+                    user_id="user-123",
+                    search_query="test",
+                    include_marketplace=False,
+                )
+
+        # Marketplace should not be called
+        mock_marketplace.assert_not_called()
+        assert len(result) == 1
+
+    @pytest.mark.asyncio
+    async def test_skips_marketplace_when_no_search_query(self):
+        """Test that marketplace is not searched without a search query."""
+        library_agents = [
+            {
+                "graph_id": "lib-123",
+                "graph_version": 1,
+                "name": "Library Agent",
+                "description": "From library",
+                "input_schema": {},
+                "output_schema": {},
+            }
+        ]
+
+        with patch.object(
+            core,
+            "get_library_agents_for_generation",
+            new_callable=AsyncMock,
+            return_value=library_agents,
+        ):
+            with patch.object(
+                core,
+                "search_marketplace_agents_for_generation",
+                new_callable=AsyncMock,
+            ) as mock_marketplace:
+                result = await core.get_all_relevant_agents_for_generation(
+                    user_id="user-123",
+                    search_query=None,  # No search query
+                    include_marketplace=True,
+                )
+
+        # Marketplace should not be called without search query
+        mock_marketplace.assert_not_called()
+        assert len(result) == 1
+
+
+class TestExtractSearchTermsFromSteps:
+    """Test extract_search_terms_from_steps function."""
+
+    def test_extracts_terms_from_instructions_type(self):
+        """Test extraction from valid instructions decomposition result."""
+        decomposition_result = {
+            "type": "instructions",
+            "steps": [
+                {
+                    "description": "Send an email notification",
+                    "block_name": "GmailSendBlock",
+                },
+                {"description": "Fetch weather data", "action": "Get weather API"},
+            ],
+        }
+
+        result = core.extract_search_terms_from_steps(decomposition_result)
+
+        assert "Send an email notification" in result
+        assert "GmailSendBlock" in result
+        assert "Fetch weather data" in result
+        assert "Get weather API" in result
+
+    def test_returns_empty_for_non_instructions_type(self):
+        """Test that non-instructions types return empty list."""
+        decomposition_result = {
+            "type": "clarifying_questions",
+            "questions": [{"question": "What email?"}],
+        }
+
+        result = core.extract_search_terms_from_steps(decomposition_result)
+
+        assert result == []
+
+    def test_deduplicates_terms_case_insensitively(self):
+        """Test that duplicate terms are removed (case-insensitive)."""
+        decomposition_result = {
+            "type": "instructions",
+            "steps": [
+                {"description": "Send Email", "name": "send email"},
+                {"description": "Other task"},
+            ],
+        }
+
+        result = core.extract_search_terms_from_steps(decomposition_result)
+
+        # Should only have one "send email" variant
+        email_terms = [t for t in result if "email" in t.lower()]
+        assert len(email_terms) == 1
+
+    def test_filters_short_terms(self):
+        """Test that terms with 3 or fewer characters are filtered out."""
+        decomposition_result = {
+            "type": "instructions",
+            "steps": [
+                {"description": "ab", "action": "xyz"},  # Both too short
+                {"description": "Valid term here"},
+            ],
+        }
+
+        result = core.extract_search_terms_from_steps(decomposition_result)
+
+        assert "ab" not in result
+        assert "xyz" not in result
+        assert "Valid term here" in result
+
+    def test_handles_empty_steps(self):
+        """Test handling of empty steps list."""
+        decomposition_result = {
+            "type": "instructions",
+            "steps": [],
+        }
+
+        result = core.extract_search_terms_from_steps(decomposition_result)
+
+        assert result == []
+
+
+class TestEnrichLibraryAgentsFromSteps:
+    """Test enrich_library_agents_from_steps function."""
+
+    @pytest.mark.asyncio
+    async def test_enriches_with_additional_agents(self):
+        """Test that additional agents are found based on steps."""
+        existing_agents = [
+            {
+                "graph_id": "existing-123",
+                "graph_version": 1,
+                "name": "Existing Agent",
+                "description": "Already fetched",
+                "input_schema": {},
+                "output_schema": {},
+            }
+        ]
+
+        additional_agents = [
+            {
+                "graph_id": "new-456",
+                "graph_version": 1,
+                "name": "Email Agent",
+                "description": "For sending emails",
+                "input_schema": {},
+                "output_schema": {},
+            }
+        ]
+
+        decomposition_result = {
+            "type": "instructions",
+            "steps": [
+                {"description": "Send email notification"},
+            ],
+        }
+
+        with patch.object(
+            core,
+            "get_all_relevant_agents_for_generation",
+            new_callable=AsyncMock,
+            return_value=additional_agents,
+        ):
+            result = await core.enrich_library_agents_from_steps(
+                user_id="user-123",
+                decomposition_result=decomposition_result,
+                existing_agents=existing_agents,
+            )
+
+        # Should have both existing and new agents
+        assert len(result) == 2
+        names = [a["name"] for a in result]
+        assert "Existing Agent" in names
+        assert "Email Agent" in names
+
+    @pytest.mark.asyncio
+    async def test_deduplicates_by_graph_id(self):
+        """Test that agents with same graph_id are not duplicated."""
+        existing_agents = [
+            {
+                "graph_id": "agent-123",
+                "graph_version": 1,
+                "name": "Existing Agent",
+                "description": "Already fetched",
+                "input_schema": {},
+                "output_schema": {},
+            }
+        ]
+
+        # Additional search returns same agent
+        additional_agents = [
+            {
+                "graph_id": "agent-123",  # Same ID
+                "graph_version": 1,
+                "name": "Existing Agent Copy",
+                "description": "Same agent different name",
+                "input_schema": {},
+                "output_schema": {},
+            }
+        ]
+
+        decomposition_result = {
+            "type": "instructions",
+            "steps": [{"description": "Some action"}],
+        }
+
+        with patch.object(
+            core,
+            "get_all_relevant_agents_for_generation",
+            new_callable=AsyncMock,
+            return_value=additional_agents,
+        ):
+            result = await core.enrich_library_agents_from_steps(
+                user_id="user-123",
+                decomposition_result=decomposition_result,
+                existing_agents=existing_agents,
+            )
+
+        # Should not duplicate
+        assert len(result) == 1
+
+    @pytest.mark.asyncio
+    async def test_deduplicates_by_name(self):
+        """Test that agents with same name are not duplicated."""
+        existing_agents = [
+            {
+                "graph_id": "agent-123",
+                "graph_version": 1,
+                "name": "Email Agent",
+                "description": "Already fetched",
+                "input_schema": {},
+                "output_schema": {},
+            }
+        ]
+
+        # Additional search returns agent with same name but different ID
+        additional_agents = [
+            {
+                "graph_id": "agent-456",  # Different ID
+                "graph_version": 1,
+                "name": "Email Agent",  # Same name
+                "description": "Different agent same name",
+                "input_schema": {},
+                "output_schema": {},
+            }
+        ]
+
+        decomposition_result = {
+            "type": "instructions",
+            "steps": [{"description": "Send email"}],
+        }
+
+        with patch.object(
+            core,
+            "get_all_relevant_agents_for_generation",
+            new_callable=AsyncMock,
+            return_value=additional_agents,
+        ):
+            result = await core.enrich_library_agents_from_steps(
+                user_id="user-123",
+                decomposition_result=decomposition_result,
+                existing_agents=existing_agents,
+            )
+
+        # Should not duplicate by name
+        assert len(result) == 1
+        assert result[0].get("graph_id") == "agent-123"  # Original kept
+
+    @pytest.mark.asyncio
+    async def test_returns_existing_when_no_steps(self):
+        """Test that existing agents are returned when no search terms extracted."""
+        existing_agents = [
+            {
+                "graph_id": "existing-123",
+                "graph_version": 1,
+                "name": "Existing Agent",
+                "description": "Already fetched",
+                "input_schema": {},
+                "output_schema": {},
+            }
+        ]
+
+        decomposition_result = {
+            "type": "clarifying_questions",  # Not instructions type
+            "questions": [],
+        }
+
+        result = await core.enrich_library_agents_from_steps(
+            user_id="user-123",
+            decomposition_result=decomposition_result,
+            existing_agents=existing_agents,
+        )
+
+        # Should return existing unchanged
+        assert result == existing_agents
+
+    @pytest.mark.asyncio
+    async def test_limits_search_terms_to_three(self):
+        """Test that only first 3 search terms are used."""
+        existing_agents = []
+
+        decomposition_result = {
+            "type": "instructions",
+            "steps": [
+                {"description": "First action"},
+                {"description": "Second action"},
+                {"description": "Third action"},
+                {"description": "Fourth action"},
+                {"description": "Fifth action"},
+            ],
+        }
+
+        call_count = 0
+
+        async def mock_get_agents(*args, **kwargs):
+            nonlocal call_count
+            call_count += 1
+            return []
+
+        with patch.object(
+            core,
+            "get_all_relevant_agents_for_generation",
+            side_effect=mock_get_agents,
+        ):
+            await core.enrich_library_agents_from_steps(
+                user_id="user-123",
+                decomposition_result=decomposition_result,
+                existing_agents=existing_agents,
+            )
+
+        # Should only make 3 calls (limited to first 3 terms)
+        assert call_count == 3
+
+
+class TestExtractUuidsFromText:
+    """Test extract_uuids_from_text function."""
+
+    def test_extracts_single_uuid(self):
+        """Test extraction of a single UUID from text."""
+        text = "Use my agent 46631191-e8a8-486f-ad90-84f89738321d for this task"
+        result = core.extract_uuids_from_text(text)
+        assert len(result) == 1
+        assert "46631191-e8a8-486f-ad90-84f89738321d" in result
+
+    def test_extracts_multiple_uuids(self):
+        """Test extraction of multiple UUIDs from text."""
+        text = (
+            "Combine agents 11111111-1111-4111-8111-111111111111 "
+            "and 22222222-2222-4222-9222-222222222222"
+        )
+        result = core.extract_uuids_from_text(text)
+        assert len(result) == 2
+        assert "11111111-1111-4111-8111-111111111111" in result
+        assert "22222222-2222-4222-9222-222222222222" in result
+
+    def test_deduplicates_uuids(self):
+        """Test that duplicate UUIDs are deduplicated."""
+        text = (
+            "Use 46631191-e8a8-486f-ad90-84f89738321d twice: "
+            "46631191-e8a8-486f-ad90-84f89738321d"
+        )
+        result = core.extract_uuids_from_text(text)
+        assert len(result) == 1
+
+    def test_normalizes_to_lowercase(self):
+        """Test that UUIDs are normalized to lowercase."""
+        text = "Use 46631191-E8A8-486F-AD90-84F89738321D"
+        result = core.extract_uuids_from_text(text)
+        assert result[0] == "46631191-e8a8-486f-ad90-84f89738321d"
+
+    def test_returns_empty_for_no_uuids(self):
+        """Test that empty list is returned when no UUIDs found."""
+        text = "Create an email agent that sends notifications"
+        result = core.extract_uuids_from_text(text)
+        assert result == []
+
+    def test_ignores_invalid_uuids(self):
+        """Test that invalid UUID-like strings are ignored."""
+        text = "Not a valid UUID: 12345678-1234-1234-1234-123456789abc"
+        result = core.extract_uuids_from_text(text)
+        # UUID v4 requires specific patterns (4 in third group, 8/9/a/b in fourth)
+        assert len(result) == 0
+
+
+class TestGetLibraryAgentById:
+    """Test get_library_agent_by_id function (and its alias get_library_agent_by_graph_id)."""
+
+    @pytest.mark.asyncio
+    async def test_returns_agent_when_found_by_graph_id(self):
+        """Test that agent is returned when found by graph_id."""
+        mock_agent = MagicMock()
+        mock_agent.graph_id = "agent-123"
+        mock_agent.graph_version = 1
+        mock_agent.name = "Test Agent"
+        mock_agent.description = "Test description"
+        mock_agent.input_schema = {"properties": {}}
+        mock_agent.output_schema = {"properties": {}}
+
+        with patch.object(
+            core.library_db,
+            "get_library_agent_by_graph_id",
+            new_callable=AsyncMock,
+            return_value=mock_agent,
+        ):
+            result = await core.get_library_agent_by_id("user-123", "agent-123")
+
+        assert result is not None
+        assert result["graph_id"] == "agent-123"
+        assert result["name"] == "Test Agent"
+
+    @pytest.mark.asyncio
+    async def test_falls_back_to_library_agent_id(self):
+        """Test that lookup falls back to library agent ID when graph_id not found."""
+        mock_agent = MagicMock()
+        mock_agent.graph_id = "graph-456"  # Different from the lookup ID
+        mock_agent.graph_version = 1
+        mock_agent.name = "Library Agent"
+        mock_agent.description = "Found by library ID"
+        mock_agent.input_schema = {"properties": {}}
+        mock_agent.output_schema = {"properties": {}}
+
+        with (
+            patch.object(
+                core.library_db,
+                "get_library_agent_by_graph_id",
+                new_callable=AsyncMock,
+                return_value=None,  # Not found by graph_id
+            ),
+            patch.object(
+                core.library_db,
+                "get_library_agent",
+                new_callable=AsyncMock,
+                return_value=mock_agent,  # Found by library ID
+            ),
+        ):
+            result = await core.get_library_agent_by_id("user-123", "library-id-123")
+
+        assert result is not None
+        assert result["graph_id"] == "graph-456"
+        assert result["name"] == "Library Agent"
+
+    @pytest.mark.asyncio
+    async def test_returns_none_when_not_found_by_either_method(self):
+        """Test that None is returned when agent not found by either method."""
+        with (
+            patch.object(
+                core.library_db,
+                "get_library_agent_by_graph_id",
+                new_callable=AsyncMock,
+                return_value=None,
+            ),
+            patch.object(
+                core.library_db,
+                "get_library_agent",
+                new_callable=AsyncMock,
+                side_effect=core.NotFoundError("Not found"),
+            ),
+        ):
+            result = await core.get_library_agent_by_id("user-123", "nonexistent")
+
+        assert result is None
+
+    @pytest.mark.asyncio
+    async def test_returns_none_on_exception(self):
+        """Test that None is returned when exception occurs in both lookups."""
+        with (
+            patch.object(
+                core.library_db,
+                "get_library_agent_by_graph_id",
+                new_callable=AsyncMock,
+                side_effect=Exception("Database error"),
+            ),
+            patch.object(
+                core.library_db,
+                "get_library_agent",
+                new_callable=AsyncMock,
+                side_effect=Exception("Database error"),
+            ),
+        ):
+            result = await core.get_library_agent_by_id("user-123", "agent-123")
+
+        assert result is None
+
+    @pytest.mark.asyncio
+    async def test_alias_works(self):
+        """Test that get_library_agent_by_graph_id is an alias for get_library_agent_by_id."""
+        assert core.get_library_agent_by_graph_id is core.get_library_agent_by_id
+
+
+class TestGetAllRelevantAgentsWithUuids:
+    """Test UUID extraction in get_all_relevant_agents_for_generation."""
+
+    @pytest.mark.asyncio
+    async def test_fetches_explicitly_mentioned_agents(self):
+        """Test that agents mentioned by UUID are fetched directly."""
+        mock_agent = MagicMock()
+        mock_agent.graph_id = "46631191-e8a8-486f-ad90-84f89738321d"
+        mock_agent.graph_version = 1
+        mock_agent.name = "Mentioned Agent"
+        mock_agent.description = "Explicitly mentioned"
+        mock_agent.input_schema = {}
+        mock_agent.output_schema = {}
+
+        mock_response = MagicMock()
+        mock_response.agents = []
+
+        with (
+            patch.object(
+                core.library_db,
+                "get_library_agent_by_graph_id",
+                new_callable=AsyncMock,
+                return_value=mock_agent,
+            ),
+            patch.object(
+                core.library_db,
+                "list_library_agents",
+                new_callable=AsyncMock,
+                return_value=mock_response,
+            ),
+        ):
+            result = await core.get_all_relevant_agents_for_generation(
+                user_id="user-123",
+                search_query="Use agent 46631191-e8a8-486f-ad90-84f89738321d",
+                include_marketplace=False,
+            )
+
+        assert len(result) == 1
+        assert result[0].get("graph_id") == "46631191-e8a8-486f-ad90-84f89738321d"
+        assert result[0].get("name") == "Mentioned Agent"
+
+
+if __name__ == "__main__":
+    pytest.main([__file__, "-v"])

autogpt_platform/backend/test/agent_generator/test_service.py

@@ -433,5 +433,139 @@ class TestGetBlocksExternal:
         assert result is None
 
 
+class TestLibraryAgentsPassthrough:
+    """Test that library_agents are passed correctly in all requests."""
+
+    def setup_method(self):
+        """Reset client singleton before each test."""
+        service._settings = None
+        service._client = None
+
+    @pytest.mark.asyncio
+    async def test_decompose_goal_passes_library_agents(self):
+        """Test that library_agents are included in decompose goal payload."""
+        library_agents = [
+            {
+                "graph_id": "agent-123",
+                "graph_version": 1,
+                "name": "Email Sender",
+                "description": "Sends emails",
+                "input_schema": {"properties": {"to": {"type": "string"}}},
+                "output_schema": {"properties": {"sent": {"type": "boolean"}}},
+            },
+        ]
+
+        mock_response = MagicMock()
+        mock_response.json.return_value = {
+            "success": True,
+            "type": "instructions",
+            "steps": ["Step 1"],
+        }
+        mock_response.raise_for_status = MagicMock()
+
+        mock_client = AsyncMock()
+        mock_client.post.return_value = mock_response
+
+        with patch.object(service, "_get_client", return_value=mock_client):
+            await service.decompose_goal_external(
+                "Send an email",
+                library_agents=library_agents,
+            )
+
+        # Verify library_agents was passed in the payload
+        call_args = mock_client.post.call_args
+        assert call_args[1]["json"]["library_agents"] == library_agents
+
+    @pytest.mark.asyncio
+    async def test_generate_agent_passes_library_agents(self):
+        """Test that library_agents are included in generate agent payload."""
+        library_agents = [
+            {
+                "graph_id": "agent-456",
+                "graph_version": 2,
+                "name": "Data Fetcher",
+                "description": "Fetches data from API",
+                "input_schema": {"properties": {"url": {"type": "string"}}},
+                "output_schema": {"properties": {"data": {"type": "object"}}},
+            },
+        ]
+
+        mock_response = MagicMock()
+        mock_response.json.return_value = {
+            "success": True,
+            "agent_json": {"name": "Test Agent", "nodes": []},
+        }
+        mock_response.raise_for_status = MagicMock()
+
+        mock_client = AsyncMock()
+        mock_client.post.return_value = mock_response
+
+        with patch.object(service, "_get_client", return_value=mock_client):
+            await service.generate_agent_external(
+                {"steps": ["Step 1"]},
+                library_agents=library_agents,
+            )
+
+        # Verify library_agents was passed in the payload
+        call_args = mock_client.post.call_args
+        assert call_args[1]["json"]["library_agents"] == library_agents
+
+    @pytest.mark.asyncio
+    async def test_generate_agent_patch_passes_library_agents(self):
+        """Test that library_agents are included in patch generation payload."""
+        library_agents = [
+            {
+                "graph_id": "agent-789",
+                "graph_version": 1,
+                "name": "Slack Notifier",
+                "description": "Sends Slack messages",
+                "input_schema": {"properties": {"message": {"type": "string"}}},
+                "output_schema": {"properties": {"success": {"type": "boolean"}}},
+            },
+        ]
+
+        mock_response = MagicMock()
+        mock_response.json.return_value = {
+            "success": True,
+            "agent_json": {"name": "Updated Agent", "nodes": []},
+        }
+        mock_response.raise_for_status = MagicMock()
+
+        mock_client = AsyncMock()
+        mock_client.post.return_value = mock_response
+
+        with patch.object(service, "_get_client", return_value=mock_client):
+            await service.generate_agent_patch_external(
+                "Add error handling",
+                {"name": "Original Agent", "nodes": []},
+                library_agents=library_agents,
+            )
+
+        # Verify library_agents was passed in the payload
+        call_args = mock_client.post.call_args
+        assert call_args[1]["json"]["library_agents"] == library_agents
+
+    @pytest.mark.asyncio
+    async def test_decompose_goal_without_library_agents(self):
+        """Test that decompose goal works without library_agents."""
+        mock_response = MagicMock()
+        mock_response.json.return_value = {
+            "success": True,
+            "type": "instructions",
+            "steps": ["Step 1"],
+        }
+        mock_response.raise_for_status = MagicMock()
+
+        mock_client = AsyncMock()
+        mock_client.post.return_value = mock_response
+
+        with patch.object(service, "_get_client", return_value=mock_client):
+            await service.decompose_goal_external("Build a workflow")
+
+        # Verify library_agents was NOT passed when not provided
+        call_args = mock_client.post.call_args
+        assert "library_agents" not in call_args[1]["json"]
+
+
 if __name__ == "__main__":
     pytest.main([__file__, "-v"])

autogpt_platform/backend/test/e2e_test_data.py

@@ -43,19 +43,24 @@ faker = Faker()
 # Constants for data generation limits (reduced for E2E tests)
 NUM_USERS = 15
 NUM_AGENT_BLOCKS = 30
-MIN_GRAPHS_PER_USER = 15
+MIN_GRAPHS_PER_USER = 25
-MAX_GRAPHS_PER_USER = 15
+MAX_GRAPHS_PER_USER = 25
 MIN_NODES_PER_GRAPH = 3
 MAX_NODES_PER_GRAPH = 6
 MIN_PRESETS_PER_USER = 2
 MAX_PRESETS_PER_USER = 3
-MIN_AGENTS_PER_USER = 15
+MIN_AGENTS_PER_USER = 25
-MAX_AGENTS_PER_USER = 15
+MAX_AGENTS_PER_USER = 25
 MIN_EXECUTIONS_PER_GRAPH = 2
 MAX_EXECUTIONS_PER_GRAPH = 8
 MIN_REVIEWS_PER_VERSION = 2
 MAX_REVIEWS_PER_VERSION = 5
 
+# Guaranteed minimums for marketplace tests (deterministic)
+GUARANTEED_FEATURED_AGENTS = 8
+GUARANTEED_FEATURED_CREATORS = 5
+GUARANTEED_TOP_AGENTS = 10
+
 
 def get_image():
     """Generate a consistent image URL using picsum.photos service."""
@@ -385,7 +390,7 @@ class TestDataCreator:
 
         library_agents = []
         for user in self.users:
-            num_agents = 10  # Create exactly 10 agents per user
+            num_agents = random.randint(MIN_AGENTS_PER_USER, MAX_AGENTS_PER_USER)
 
             # Get available graphs for this user
             user_graphs = [
@@ -507,14 +512,17 @@ class TestDataCreator:
             existing_profiles, min(num_creators, len(existing_profiles))
         )
 
-        # Mark about 50% of creators as featured (more for testing)
+        # Guarantee at least GUARANTEED_FEATURED_CREATORS featured creators
-        num_featured = max(2, int(num_creators * 0.5))
+        num_featured = max(GUARANTEED_FEATURED_CREATORS, int(num_creators * 0.5))
         num_featured = min(
             num_featured, len(selected_profiles)
         )  # Don't exceed available profiles
         featured_profile_ids = set(
             random.sample([p.id for p in selected_profiles], num_featured)
         )
+        print(
+            f"🎯 Creating {num_featured} featured creators (min: {GUARANTEED_FEATURED_CREATORS})"
+        )
 
         for profile in selected_profiles:
             try:
@@ -545,21 +553,25 @@ class TestDataCreator:
         return profiles
 
     async def create_test_store_submissions(self) -> List[Dict[str, Any]]:
-        """Create test store submissions using the API function."""
+        """Create test store submissions using the API function.
+
+        DETERMINISTIC: Guarantees minimum featured agents for E2E tests.
+        """
         print("Creating test store submissions...")
 
         submissions = []
         approved_submissions = []
+        featured_count = 0
+        submission_counter = 0
 
-        # Create a special test submission for test123@gmail.com
+        # Create a special test submission for test123@gmail.com (ALWAYS approved + featured)
         test_user = next(
             (user for user in self.users if user["email"] == "test123@gmail.com"), None
         )
-        if test_user:
+        if test_user and self.agent_graphs:
-            # Special test data for consistent testing
             test_submission_data = {
                 "user_id": test_user["id"],
-                "agent_id": self.agent_graphs[0]["id"],  # Use first available graph
+                "agent_id": self.agent_graphs[0]["id"],
                 "agent_version": 1,
                 "slug": "test-agent-submission",
                 "name": "Test Agent Submission",
@@ -580,37 +592,24 @@ class TestDataCreator:
                 submissions.append(test_submission.model_dump())
                 print("✅ Created special test store submission for test123@gmail.com")
 
-                # Randomly approve, reject, or leave pending the test submission
+                # ALWAYS approve and feature the test submission
                 if test_submission.store_listing_version_id:
-                    random_value = random.random()
+                    approved_submission = await review_store_submission(
-                    if random_value < 0.4:  # 40% chance to approve
+                        store_listing_version_id=test_submission.store_listing_version_id,
-                        approved_submission = await review_store_submission(
+                        is_approved=True,
-                            store_listing_version_id=test_submission.store_listing_version_id,
+                        external_comments="Test submission approved",
-                            is_approved=True,
+                        internal_comments="Auto-approved test submission",
-                            external_comments="Test submission approved",
+                        reviewer_id=test_user["id"],
-                            internal_comments="Auto-approved test submission",
+                    )
-                            reviewer_id=test_user["id"],
+                    approved_submissions.append(approved_submission.model_dump())
-                        )
+                    print("✅ Approved test store submission")
-                        approved_submissions.append(approved_submission.model_dump())
-                        print("✅ Approved test store submission")
 
-                        # Mark approved submission as featured
+                    await prisma.storelistingversion.update(
-                        await prisma.storelistingversion.update(
+                        where={"id": test_submission.store_listing_version_id},
-                            where={"id": test_submission.store_listing_version_id},
+                        data={"isFeatured": True},
-                            data={"isFeatured": True},
+                    )
-                        )
+                    featured_count += 1
-                        print("🌟 Marked test agent as FEATURED")
+                    print("🌟 Marked test agent as FEATURED")
-                    elif random_value < 0.7:  # 30% chance to reject (40% to 70%)
-                        await review_store_submission(
-                            store_listing_version_id=test_submission.store_listing_version_id,
-                            is_approved=False,
-                            external_comments="Test submission rejected - needs improvements",
-                            internal_comments="Auto-rejected test submission for E2E testing",
-                            reviewer_id=test_user["id"],
-                        )
-                        print("❌ Rejected test store submission")
-                    else:  # 30% chance to leave pending (70% to 100%)
-                        print("⏳ Left test submission pending for review")
 
             except Exception as e:
                 print(f"Error creating test store submission: {e}")
@@ -620,7 +619,6 @@ class TestDataCreator:
 
         # Create regular submissions for all users
         for user in self.users:
-            # Get available graphs for this specific user
             user_graphs = [
                 g for g in self.agent_graphs if g.get("userId") == user["id"]
             ]
@@ -631,18 +629,17 @@ class TestDataCreator:
                 )
                 continue
 
-            # Create exactly 4 store submissions per user
             for submission_index in range(4):
                 graph = random.choice(user_graphs)
+                submission_counter += 1
 
                 try:
                     print(
-                        f"Creating store submission for user {user['id']} with graph {graph['id']} (owner: {graph.get('userId')})"
+                        f"Creating store submission for user {user['id']} with graph {graph['id']}"
                     )
 
-                    # Use the API function to create store submission with correct parameters
                     submission = await create_store_submission(
-                        user_id=user["id"],  # Must match graph's userId
+                        user_id=user["id"],
                         agent_id=graph["id"],
                         agent_version=graph.get("version", 1),
                         slug=faker.slug(),
@@ -651,22 +648,24 @@ class TestDataCreator:
                         video_url=get_video_url() if random.random() < 0.3 else None,
                         image_urls=[get_image() for _ in range(3)],
                         description=faker.text(),
-                        categories=[
+                        categories=[get_category()],
-                            get_category()
-                        ],  # Single category from predefined list
                         changes_summary="Initial E2E test submission",
                     )
                     submissions.append(submission.model_dump())
                     print(f"✅ Created store submission: {submission.name}")
 
-                    # Randomly approve, reject, or leave pending the submission
                     if submission.store_listing_version_id:
-                        random_value = random.random()
+                        # DETERMINISTIC: First N submissions are always approved
-                        if random_value < 0.4:  # 40% chance to approve
+                        # First GUARANTEED_FEATURED_AGENTS of those are always featured
-                            try:
+                        should_approve = (
-                                # Pick a random user as the reviewer (admin)
+                            submission_counter <= GUARANTEED_TOP_AGENTS
-                                reviewer_id = random.choice(self.users)["id"]
+                            or random.random() < 0.4
+                        )
+                        should_feature = featured_count < GUARANTEED_FEATURED_AGENTS
 
+                        if should_approve:
+                            try:
+                                reviewer_id = random.choice(self.users)["id"]
                                 approved_submission = await review_store_submission(
                                     store_listing_version_id=submission.store_listing_version_id,
                                     is_approved=True,
@@ -681,16 +680,7 @@ class TestDataCreator:
                                     f"✅ Approved store submission: {submission.name}"
                                 )
 
-                                # Mark some agents as featured during creation (30% chance)
+                                if should_feature:
-                                # More likely for creators and first submissions
-                                is_creator = user["id"] in [
-                                    p.get("userId") for p in self.profiles
-                                ]
-                                feature_chance = (
-                                    0.5 if is_creator else 0.2
-                                )  # 50% for creators, 20% for others
-
-                                if random.random() < feature_chance:
                                     try:
                                         await prisma.storelistingversion.update(
                                             where={
@@ -698,8 +688,25 @@ class TestDataCreator:
                                             },
                                             data={"isFeatured": True},
                                         )
+                                        featured_count += 1
                                         print(
-                                            f"🌟 Marked agent as FEATURED: {submission.name}"
+                                            f"🌟 Marked agent as FEATURED ({featured_count}/{GUARANTEED_FEATURED_AGENTS}): {submission.name}"
+                                        )
+                                    except Exception as e:
+                                        print(
+                                            f"Warning: Could not mark submission as featured: {e}"
+                                        )
+                                elif random.random() < 0.2:
+                                    try:
+                                        await prisma.storelistingversion.update(
+                                            where={
+                                                "id": submission.store_listing_version_id
+                                            },
+                                            data={"isFeatured": True},
+                                        )
+                                        featured_count += 1
+                                        print(
+                                            f"🌟 Marked agent as FEATURED (bonus): {submission.name}"
                                         )
                                     except Exception as e:
                                         print(
@@ -710,11 +717,9 @@ class TestDataCreator:
                                 print(
                                     f"Warning: Could not approve submission {submission.name}: {e}"
                                 )
-                        elif random_value < 0.7:  # 30% chance to reject (40% to 70%)
+                        elif random.random() < 0.5:
                             try:
-                                # Pick a random user as the reviewer (admin)
                                 reviewer_id = random.choice(self.users)["id"]
-
                                 await review_store_submission(
                                     store_listing_version_id=submission.store_listing_version_id,
                                     is_approved=False,
@@ -729,7 +734,7 @@ class TestDataCreator:
                                 print(
                                     f"Warning: Could not reject submission {submission.name}: {e}"
                                 )
-                        else:  # 30% chance to leave pending (70% to 100%)
+                        else:
                             print(
                                 f"⏳ Left submission pending for review: {submission.name}"
                             )
@@ -743,9 +748,13 @@ class TestDataCreator:
                     traceback.print_exc()
                     continue
 
+        print("\n📊 Store Submissions Summary:")
+        print(f"   Created: {len(submissions)}")
+        print(f"   Approved: {len(approved_submissions)}")
         print(
-            f"Created {len(submissions)} store submissions, approved {len(approved_submissions)}"
+            f"   Featured: {featured_count} (guaranteed min: {GUARANTEED_FEATURED_AGENTS})"
         )
+
         self.store_submissions = submissions
         return submissions
 
@@ -825,12 +834,15 @@ class TestDataCreator:
         print(f"✅ Agent blocks available: {len(self.agent_blocks)}")
         print(f"✅ Agent graphs created: {len(self.agent_graphs)}")
         print(f"✅ Library agents created: {len(self.library_agents)}")
-        print(f"✅ Creator profiles updated: {len(self.profiles)} (some featured)")
+        print(f"✅ Creator profiles updated: {len(self.profiles)}")
-        print(
+        print(f"✅ Store submissions created: {len(self.store_submissions)}")
-            f"✅ Store submissions created: {len(self.store_submissions)} (some marked as featured during creation)"
-        )
         print(f"✅ API keys created: {len(self.api_keys)}")
         print(f"✅ Presets created: {len(self.presets)}")
+        print("\n🎯 Deterministic Guarantees:")
+        print(f"   • Featured agents: >= {GUARANTEED_FEATURED_AGENTS}")
+        print(f"   • Featured creators: >= {GUARANTEED_FEATURED_CREATORS}")
+        print(f"   • Top agents (approved): >= {GUARANTEED_TOP_AGENTS}")
+        print(f"   • Library agents per user: >= {MIN_AGENTS_PER_USER}")
         print("\n🚀 Your E2E test database is ready to use!")
 
 

autogpt_platform/frontend/src/app/(platform)/build/components/legacy-builder/CustomNode/CustomNode.tsx

@@ -528,6 +528,10 @@ export const CustomNode = React.memo(
           const getInputPropKey = (key: string) =>
             nodeType == BlockUIType.AGENT ? `inputs.${key}` : key;
 
+          // For AGENT blocks, handle IDs need the inputs_#_ prefix to match link sink_names
+          const getHandleKey = (key: string) =>
+            nodeType == BlockUIType.AGENT ? `inputs_#_${key}` : key;
+
           return keys.map(([propKey, propSchema]) => {
             const isRequired = data.inputSchema.required?.includes(propKey);
             const isAdvanced = propSchema.advanced;
@@ -544,7 +548,8 @@ export const CustomNode = React.memo(
               !propKey.endsWith("_credentials") &&
               // For OUTPUT blocks, only show the 'value' (hides 'name') input connection handle
               !(nodeType == BlockUIType.OUTPUT && propKey == "name");
-            const isConnected = isInputHandleConnected(propKey);
+            const handleKey = getHandleKey(propKey);
+            const isConnected = isInputHandleConnected(handleKey);
 
             return (
               !isHidden &&
@@ -562,12 +567,12 @@ export const CustomNode = React.memo(
                     propSchema.discriminator
                   ) ? (
                     <NodeHandle
-                      keyName={propKey}
+                      keyName={handleKey}
                       isConnected={isConnected}
                       isRequired={isRequired}
                       schema={propSchema}
                       side="left"
-                      isBroken={isInputHandleBroken(propKey)}
+                      isBroken={isInputHandleBroken(handleKey)}
                     />
                   ) : (
                     propKey !== "credentials" &&
@@ -857,7 +862,7 @@ export const CustomNode = React.memo(
     })();
 
     const hasAdvancedFields =
-      data.inputSchema &&
+      data.inputSchema?.properties &&
       Object.entries(data.inputSchema.properties).some(([key, value]) => {
         return (
           value.advanced === true && !data.inputSchema.required?.includes(key)

autogpt_platform/frontend/src/app/api/openapi.json

@@ -7981,6 +7981,25 @@
             ]
           },
           "new_output": { "type": "boolean", "title": "New Output" },
+          "execution_count": {
+            "type": "integer",
+            "title": "Execution Count",
+            "default": 0
+          },
+          "success_rate": {
+            "anyOf": [{ "type": "number" }, { "type": "null" }],
+            "title": "Success Rate"
+          },
+          "avg_correctness_score": {
+            "anyOf": [{ "type": "number" }, { "type": "null" }],
+            "title": "Avg Correctness Score"
+          },
+          "recent_executions": {
+            "items": { "$ref": "#/components/schemas/RecentExecution" },
+            "type": "array",
+            "title": "Recent Executions",
+            "description": "List of recent executions with status, score, and summary"
+          },
           "can_access_graph": {
             "type": "boolean",
             "title": "Can Access Graph"
@@ -9374,6 +9393,23 @@
         "required": ["providers", "pagination"],
         "title": "ProviderResponse"
       },
+      "RecentExecution": {
+        "properties": {
+          "status": { "type": "string", "title": "Status" },
+          "correctness_score": {
+            "anyOf": [{ "type": "number" }, { "type": "null" }],
+            "title": "Correctness Score"
+          },
+          "activity_summary": {
+            "anyOf": [{ "type": "string" }, { "type": "null" }],
+            "title": "Activity Summary"
+          }
+        },
+        "type": "object",
+        "required": ["status"],
+        "title": "RecentExecution",
+        "description": "Summary of a recent execution for quality assessment.\n\nUsed by the LLM to understand the agent's recent performance with specific examples\nrather than just aggregate statistics."
+      },
       "RefundRequest": {
         "properties": {
           "id": { "type": "string", "title": "Id" },

autogpt_platform/frontend/src/components/contextual/Chat/components/ChatInput/ChatInput.tsx

@@ -57,6 +57,7 @@ export function ChatInput({
     isStreaming,
     value,
     baseHandleKeyDown,
+    inputId,
   });
 
   return (

autogpt_platform/frontend/src/components/contextual/Chat/components/ChatInput/useVoiceRecording.ts

@@ -15,6 +15,7 @@ interface Args {
   isStreaming?: boolean;
   value: string;
   baseHandleKeyDown: (event: KeyboardEvent<HTMLTextAreaElement>) => void;
+  inputId?: string;
 }
 
 export function useVoiceRecording({
@@ -23,6 +24,7 @@ export function useVoiceRecording({
   isStreaming = false,
   value,
   baseHandleKeyDown,
+  inputId,
 }: Args) {
   const [isRecording, setIsRecording] = useState(false);
   const [isTranscribing, setIsTranscribing] = useState(false);
@@ -103,7 +105,7 @@ export function useVoiceRecording({
         setIsTranscribing(false);
       }
     },
-    [handleTranscription],
+    [handleTranscription, inputId],
   );
 
   const stopRecording = useCallback(() => {
@@ -201,6 +203,15 @@ export function useVoiceRecording({
     }
   }, [error, toast]);
 
+  useEffect(() => {
+    if (!isTranscribing && inputId) {
+      const inputElement = document.getElementById(inputId);
+      if (inputElement) {
+        inputElement.focus();
+      }
+    }
+  }, [isTranscribing, inputId]);
+
   const handleKeyDown = useCallback(
     (event: KeyboardEvent<HTMLTextAreaElement>) => {
       if (event.key === " " && !value.trim() && !isTranscribing) {

autogpt_platform/frontend/src/components/contextual/Chat/components/ChatMessage/ChatMessage.tsx

@@ -156,11 +156,20 @@ export function ChatMessage({
   }
 
   if (isClarificationNeeded && message.type === "clarification_needed") {
+    // Check if user already replied after this clarification (answered)
+    const hasUserReplyAfter =
+      index >= 0 &&
+      messages
+        .slice(index + 1)
+        .some((m) => m.type === "message" && m.role === "user");
+
     return (
       <ClarificationQuestionsWidget
         questions={message.questions}
         message={message.message}
+        sessionId={message.sessionId}
         onSubmitAnswers={handleClarificationAnswers}
+        isAnswered={hasUserReplyAfter}
         className={className}
       />
     );

autogpt_platform/frontend/src/components/contextual/Chat/components/ClarificationQuestionsWidget/ClarificationQuestionsWidget.tsx

@@ -6,7 +6,7 @@ import { Input } from "@/components/atoms/Input/Input";
 import { Text } from "@/components/atoms/Text/Text";
 import { cn } from "@/lib/utils";
 import { CheckCircleIcon, QuestionIcon } from "@phosphor-icons/react";
-import { useState } from "react";
+import { useState, useEffect, useRef } from "react";
 
 export interface ClarifyingQuestion {
   question: string;
@@ -17,39 +17,97 @@ export interface ClarifyingQuestion {
 interface Props {
   questions: ClarifyingQuestion[];
   message: string;
+  sessionId?: string;
   onSubmitAnswers: (answers: Record<string, string>) => void;
   onCancel?: () => void;
+  isAnswered?: boolean;
   className?: string;
 }
 
+function getStorageKey(sessionId?: string): string | null {
+  if (!sessionId) return null;
+  return `clarification_answers_${sessionId}`;
+}
+
 export function ClarificationQuestionsWidget({
   questions,
   message,
+  sessionId,
   onSubmitAnswers,
   onCancel,
+  isAnswered = false,
   className,
 }: Props) {
   const [answers, setAnswers] = useState<Record<string, string>>({});
   const [isSubmitted, setIsSubmitted] = useState(false);
+  const lastSessionIdRef = useRef<string | undefined>(undefined);
+
+  useEffect(() => {
+    const storageKey = getStorageKey(sessionId);
+    if (!storageKey) {
+      setAnswers({});
+      setIsSubmitted(false);
+      lastSessionIdRef.current = sessionId;
+      return;
+    }
+
+    try {
+      const saved = localStorage.getItem(storageKey);
+      if (saved) {
+        const parsed = JSON.parse(saved) as Record<string, string>;
+        setAnswers(parsed);
+      } else {
+        setAnswers({});
+      }
+      setIsSubmitted(false);
+    } catch {
+      setAnswers({});
+      setIsSubmitted(false);
+    }
+    lastSessionIdRef.current = sessionId;
+  }, [sessionId]);
+
+  useEffect(() => {
+    if (lastSessionIdRef.current !== sessionId) {
+      return;
+    }
+    const storageKey = getStorageKey(sessionId);
+    if (!storageKey) return;
+
+    const hasAnswers = Object.values(answers).some((v) => v.trim());
+    try {
+      if (hasAnswers) {
+        localStorage.setItem(storageKey, JSON.stringify(answers));
+      } else {
+        localStorage.removeItem(storageKey);
+      }
+    } catch {}
+  }, [answers, sessionId]);
 
   function handleAnswerChange(keyword: string, value: string) {
     setAnswers((prev) => ({ ...prev, [keyword]: value }));
   }
 
   function handleSubmit() {
-    // Check if all questions are answered
     const allAnswered = questions.every((q) => answers[q.keyword]?.trim());
     if (!allAnswered) {
       return;
     }
     setIsSubmitted(true);
     onSubmitAnswers(answers);
+
+    const storageKey = getStorageKey(sessionId);
+    try {
+      if (storageKey) {
+        localStorage.removeItem(storageKey);
+      }
+    } catch {}
   }
 
   const allAnswered = questions.every((q) => answers[q.keyword]?.trim());
 
-  // Show submitted state after answers are submitted
+  // Show submitted state if answered from conversation or just submitted
-  if (isSubmitted) {
+  if (isAnswered || isSubmitted) {
     return (
       <div
         className={cn(

autogpt_platform/frontend/src/components/contextual/Chat/components/ToolResponseMessage/helpers.ts

@@ -30,9 +30,9 @@ export function getErrorMessage(result: unknown): string {
   }
   if (typeof result === "object" && result !== null) {
     const response = result as Record<string, unknown>;
-    if (response.error) return stripInternalReasoning(String(response.error));
     if (response.message)
       return stripInternalReasoning(String(response.message));
+    if (response.error) return stripInternalReasoning(String(response.error));
   }
   return "An error occurred";
 }
@@ -363,8 +363,8 @@ export function formatToolResponse(result: unknown, toolName: string): string {
 
     case "error":
       const errorMsg =
-        (response.error as string) || response.message || "An error occurred";
+        (response.message as string) || response.error || "An error occurred";
-      return `Error: ${errorMsg}`;
+      return stripInternalReasoning(String(errorMsg));
 
     case "no_results":
       const suggestions = (response.suggestions as string[]) || [];

autogpt_platform/frontend/src/tests/library.spec.ts

@@ -59,12 +59,13 @@ test.describe("Library", () => {
   });
 
   test("pagination works correctly", async ({ page }, testInfo) => {
-    test.setTimeout(testInfo.timeout * 3); // Increase timeout for pagination operations
+    test.setTimeout(testInfo.timeout * 3);
     await page.goto("/library");
 
+    const PAGE_SIZE = 20;
     const paginationResult = await libraryPage.testPagination();
 
-    if (paginationResult.initialCount >= 10) {
+    if (paginationResult.initialCount >= PAGE_SIZE) {
       expect(paginationResult.finalCount).toBeGreaterThanOrEqual(
         paginationResult.initialCount,
       );
@@ -133,7 +134,10 @@ test.describe("Library", () => {
     test.expect(clearedSearchValue).toBe("");
   });
 
-  test("pagination while searching works correctly", async ({ page }) => {
+  test("pagination while searching works correctly", async ({
+    page,
+  }, testInfo) => {
+    test.setTimeout(testInfo.timeout * 3);
     await page.goto("/library");
 
     const allAgents = await libraryPage.getAgents();
@@ -152,9 +156,10 @@ test.describe("Library", () => {
     );
     expect(matchingResults.length).toEqual(initialSearchResults.length);
 
+    const PAGE_SIZE = 20;
     const searchPaginationResult = await libraryPage.testPagination();
 
-    if (searchPaginationResult.initialCount >= 10) {
+    if (searchPaginationResult.initialCount >= PAGE_SIZE) {
       expect(searchPaginationResult.finalCount).toBeGreaterThanOrEqual(
         searchPaginationResult.initialCount,
       );

autogpt_platform/frontend/src/tests/marketplace-creator.spec.ts

@@ -69,9 +69,12 @@ test.describe("Marketplace Creator Page – Basic Functionality", () => {
       await marketplacePage.getFirstCreatorProfile(page);
     await firstCreatorProfile.click();
     await page.waitForURL("**/marketplace/creator/**");
+    await page.waitForLoadState("networkidle").catch(() => {});
+
     const firstAgent = page
       .locator('[data-testid="store-card"]:visible')
       .first();
+    await firstAgent.waitFor({ state: "visible", timeout: 30000 });
 
     await firstAgent.click();
     await page.waitForURL("**/marketplace/agent/**");

autogpt_platform/frontend/src/tests/marketplace.spec.ts

@@ -77,7 +77,6 @@ test.describe("Marketplace – Basic Functionality", () => {
 
     const firstFeaturedAgent =
       await marketplacePage.getFirstFeaturedAgent(page);
-    await firstFeaturedAgent.waitFor({ state: "visible" });
     await firstFeaturedAgent.click();
     await page.waitForURL("**/marketplace/agent/**");
     await matchesUrl(page, /\/marketplace\/agent\/.+/);
@@ -116,7 +115,15 @@ test.describe("Marketplace – Basic Functionality", () => {
     const searchTerm = page.getByText("DummyInput").first();
     await isVisible(searchTerm);
 
-    await page.waitForTimeout(10000);
+    await page.waitForLoadState("networkidle").catch(() => {});
+
+    await page
+      .waitForFunction(
+        () =>
+          document.querySelectorAll('[data-testid="store-card"]').length > 0,
+        { timeout: 15000 },
+      )
+      .catch(() => console.log("No search results appeared within timeout"));
 
     const results = await marketplacePage.getSearchResultsCount(page);
     expect(results).toBeGreaterThan(0);

autogpt_platform/frontend/src/tests/pages/library.page.ts

@@ -300,21 +300,27 @@ export class LibraryPage extends BasePage {
   async scrollToLoadMore(): Promise<void> {
     console.log(`scrolling to load more agents`);
 
-    // Get initial agent count
+    const initialCount = await this.getAgentCountByListLength();
-    const initialCount = await this.getAgentCount();
+    console.log(`Initial agent count (DOM cards): ${initialCount}`);
-    console.log(`Initial agent count: ${initialCount}`);
 
-    // Scroll down to trigger pagination
     await this.scrollToBottom();
 
-    // Wait for potential new agents to load
+    await this.page
-    await this.page.waitForTimeout(2000);
+      .waitForLoadState("networkidle", { timeout: 10000 })
+      .catch(() => console.log("Network idle timeout, continuing..."));
 
-    // Check if more agents loaded
+    await this.page
-    const newCount = await this.getAgentCount();
+      .waitForFunction(
-    console.log(`New agent count after scroll: ${newCount}`);
+        (prevCount) =>
+          document.querySelectorAll('[data-testid="library-agent-card"]')
+            .length > prevCount,
+        initialCount,
+        { timeout: 5000 },
+      )
+      .catch(() => {});
 
-    return;
+    const newCount = await this.getAgentCountByListLength();
+    console.log(`New agent count after scroll (DOM cards): ${newCount}`);
   }
 
   async testPagination(): Promise<{

autogpt_platform/frontend/src/tests/pages/marketplace.page.ts

@@ -9,6 +9,7 @@ export class MarketplacePage extends BasePage {
 
   async goto(page: Page) {
     await page.goto("/marketplace");
+    await page.waitForLoadState("networkidle").catch(() => {});
   }
 
   async getMarketplaceTitle(page: Page) {
@@ -109,16 +110,24 @@ export class MarketplacePage extends BasePage {
 
   async getFirstFeaturedAgent(page: Page) {
     const { getId } = getSelectors(page);
-    return getId("featured-store-card").first();
+    const card = getId("featured-store-card").first();
+    await card.waitFor({ state: "visible", timeout: 30000 });
+    return card;
   }
 
   async getFirstTopAgent() {
-    return this.page.locator('[data-testid="store-card"]:visible').first();
+    const card = this.page
+      .locator('[data-testid="store-card"]:visible')
+      .first();
+    await card.waitFor({ state: "visible", timeout: 30000 });
+    return card;
   }
 
   async getFirstCreatorProfile(page: Page) {
     const { getId } = getSelectors(page);
-    return getId("creator-card").first();
+    const card = getId("creator-card").first();
+    await card.waitFor({ state: "visible", timeout: 30000 });
+    return card;
   }
 
   async getSearchResultsCount(page: Page) {

docs/integrations/block-integrations/llm.md

@@ -65,7 +65,7 @@ The result routes data to yes_output or no_output, enabling intelligent branchin
 | condition | A plaintext English description of the condition to evaluate | str | Yes |
 | yes_value | (Optional) Value to output if the condition is true. If not provided, input_value will be used. | Yes Value | No |
 | no_value | (Optional) Value to output if the condition is false. If not provided, input_value will be used. | No Value | No |
-| model | The language model to use for evaluating the condition. | "o3-mini" \| "o3-2025-04-16" \| "o1" \| "o1-mini" \| "gpt-5.2-2025-12-11" \| "gpt-5.1-2025-11-13" \| "gpt-5-2025-08-07" \| "gpt-5-mini-2025-08-07" \| "gpt-5-nano-2025-08-07" \| "gpt-5-chat-latest" \| "gpt-4.1-2025-04-14" \| "gpt-4.1-mini-2025-04-14" \| "gpt-4o-mini" \| "gpt-4o" \| "gpt-4-turbo" \| "gpt-3.5-turbo" \| "claude-opus-4-1-20250805" \| "claude-opus-4-20250514" \| "claude-sonnet-4-20250514" \| "claude-opus-4-5-20251101" \| "claude-sonnet-4-5-20250929" \| "claude-haiku-4-5-20251001" \| "claude-3-7-sonnet-20250219" \| "claude-3-haiku-20240307" \| "Qwen/Qwen2.5-72B-Instruct-Turbo" \| "nvidia/llama-3.1-nemotron-70b-instruct" \| "meta-llama/Llama-3.3-70B-Instruct-Turbo" \| "meta-llama/Meta-Llama-3.1-70B-Instruct-Turbo" \| "meta-llama/Llama-3.2-3B-Instruct-Turbo" \| "llama-3.3-70b-versatile" \| "llama-3.1-8b-instant" \| "llama3.3" \| "llama3.2" \| "llama3" \| "llama3.1:405b" \| "dolphin-mistral:latest" \| "openai/gpt-oss-120b" \| "openai/gpt-oss-20b" \| "google/gemini-2.5-pro-preview-03-25" \| "google/gemini-3-pro-preview" \| "google/gemini-2.5-flash" \| "google/gemini-2.0-flash-001" \| "google/gemini-2.5-flash-lite-preview-06-17" \| "google/gemini-2.0-flash-lite-001" \| "mistralai/mistral-nemo" \| "cohere/command-r-08-2024" \| "cohere/command-r-plus-08-2024" \| "deepseek/deepseek-chat" \| "deepseek/deepseek-r1-0528" \| "perplexity/sonar" \| "perplexity/sonar-pro" \| "perplexity/sonar-deep-research" \| "nousresearch/hermes-3-llama-3.1-405b" \| "nousresearch/hermes-3-llama-3.1-70b" \| "amazon/nova-lite-v1" \| "amazon/nova-micro-v1" \| "amazon/nova-pro-v1" \| "microsoft/wizardlm-2-8x22b" \| "gryphe/mythomax-l2-13b" \| "meta-llama/llama-4-scout" \| "meta-llama/llama-4-maverick" \| "x-ai/grok-4" \| "x-ai/grok-4-fast" \| "x-ai/grok-4.1-fast" \| "x-ai/grok-code-fast-1" \| "moonshotai/kimi-k2" \| "qwen/qwen3-235b-a22b-thinking-2507" \| "qwen/qwen3-coder" \| "Llama-4-Scout-17B-16E-Instruct-FP8" \| "Llama-4-Maverick-17B-128E-Instruct-FP8" \| "Llama-3.3-8B-Instruct" \| "Llama-3.3-70B-Instruct" \| "v0-1.5-md" \| "v0-1.5-lg" \| "v0-1.0-md" | No |
+| model | The language model to use for evaluating the condition. | "o3-mini" \| "o3-2025-04-16" \| "o1" \| "o1-mini" \| "gpt-5.2-2025-12-11" \| "gpt-5.1-2025-11-13" \| "gpt-5-2025-08-07" \| "gpt-5-mini-2025-08-07" \| "gpt-5-nano-2025-08-07" \| "gpt-5-chat-latest" \| "gpt-4.1-2025-04-14" \| "gpt-4.1-mini-2025-04-14" \| "gpt-4o-mini" \| "gpt-4o" \| "gpt-4-turbo" \| "gpt-3.5-turbo" \| "claude-opus-4-1-20250805" \| "claude-opus-4-20250514" \| "claude-sonnet-4-20250514" \| "claude-opus-4-5-20251101" \| "claude-sonnet-4-5-20250929" \| "claude-haiku-4-5-20251001" \| "claude-3-haiku-20240307" \| "Qwen/Qwen2.5-72B-Instruct-Turbo" \| "nvidia/llama-3.1-nemotron-70b-instruct" \| "meta-llama/Llama-3.3-70B-Instruct-Turbo" \| "meta-llama/Meta-Llama-3.1-70B-Instruct-Turbo" \| "meta-llama/Llama-3.2-3B-Instruct-Turbo" \| "llama-3.3-70b-versatile" \| "llama-3.1-8b-instant" \| "llama3.3" \| "llama3.2" \| "llama3" \| "llama3.1:405b" \| "dolphin-mistral:latest" \| "openai/gpt-oss-120b" \| "openai/gpt-oss-20b" \| "google/gemini-2.5-pro-preview-03-25" \| "google/gemini-3-pro-preview" \| "google/gemini-2.5-flash" \| "google/gemini-2.0-flash-001" \| "google/gemini-2.5-flash-lite-preview-06-17" \| "google/gemini-2.0-flash-lite-001" \| "mistralai/mistral-nemo" \| "cohere/command-r-08-2024" \| "cohere/command-r-plus-08-2024" \| "deepseek/deepseek-chat" \| "deepseek/deepseek-r1-0528" \| "perplexity/sonar" \| "perplexity/sonar-pro" \| "perplexity/sonar-deep-research" \| "nousresearch/hermes-3-llama-3.1-405b" \| "nousresearch/hermes-3-llama-3.1-70b" \| "amazon/nova-lite-v1" \| "amazon/nova-micro-v1" \| "amazon/nova-pro-v1" \| "microsoft/wizardlm-2-8x22b" \| "gryphe/mythomax-l2-13b" \| "meta-llama/llama-4-scout" \| "meta-llama/llama-4-maverick" \| "x-ai/grok-4" \| "x-ai/grok-4-fast" \| "x-ai/grok-4.1-fast" \| "x-ai/grok-code-fast-1" \| "moonshotai/kimi-k2" \| "qwen/qwen3-235b-a22b-thinking-2507" \| "qwen/qwen3-coder" \| "Llama-4-Scout-17B-16E-Instruct-FP8" \| "Llama-4-Maverick-17B-128E-Instruct-FP8" \| "Llama-3.3-8B-Instruct" \| "Llama-3.3-70B-Instruct" \| "v0-1.5-md" \| "v0-1.5-lg" \| "v0-1.0-md" | No |
 
 ### Outputs
 
@@ -103,7 +103,7 @@ The block sends the entire conversation history to the chosen LLM, including sys
 |-------|-------------|------|----------|
 | prompt | The prompt to send to the language model. | str | No |
 | messages | List of messages in the conversation. | List[Any] | Yes |
-| model | The language model to use for the conversation. | "o3-mini" \| "o3-2025-04-16" \| "o1" \| "o1-mini" \| "gpt-5.2-2025-12-11" \| "gpt-5.1-2025-11-13" \| "gpt-5-2025-08-07" \| "gpt-5-mini-2025-08-07" \| "gpt-5-nano-2025-08-07" \| "gpt-5-chat-latest" \| "gpt-4.1-2025-04-14" \| "gpt-4.1-mini-2025-04-14" \| "gpt-4o-mini" \| "gpt-4o" \| "gpt-4-turbo" \| "gpt-3.5-turbo" \| "claude-opus-4-1-20250805" \| "claude-opus-4-20250514" \| "claude-sonnet-4-20250514" \| "claude-opus-4-5-20251101" \| "claude-sonnet-4-5-20250929" \| "claude-haiku-4-5-20251001" \| "claude-3-7-sonnet-20250219" \| "claude-3-haiku-20240307" \| "Qwen/Qwen2.5-72B-Instruct-Turbo" \| "nvidia/llama-3.1-nemotron-70b-instruct" \| "meta-llama/Llama-3.3-70B-Instruct-Turbo" \| "meta-llama/Meta-Llama-3.1-70B-Instruct-Turbo" \| "meta-llama/Llama-3.2-3B-Instruct-Turbo" \| "llama-3.3-70b-versatile" \| "llama-3.1-8b-instant" \| "llama3.3" \| "llama3.2" \| "llama3" \| "llama3.1:405b" \| "dolphin-mistral:latest" \| "openai/gpt-oss-120b" \| "openai/gpt-oss-20b" \| "google/gemini-2.5-pro-preview-03-25" \| "google/gemini-3-pro-preview" \| "google/gemini-2.5-flash" \| "google/gemini-2.0-flash-001" \| "google/gemini-2.5-flash-lite-preview-06-17" \| "google/gemini-2.0-flash-lite-001" \| "mistralai/mistral-nemo" \| "cohere/command-r-08-2024" \| "cohere/command-r-plus-08-2024" \| "deepseek/deepseek-chat" \| "deepseek/deepseek-r1-0528" \| "perplexity/sonar" \| "perplexity/sonar-pro" \| "perplexity/sonar-deep-research" \| "nousresearch/hermes-3-llama-3.1-405b" \| "nousresearch/hermes-3-llama-3.1-70b" \| "amazon/nova-lite-v1" \| "amazon/nova-micro-v1" \| "amazon/nova-pro-v1" \| "microsoft/wizardlm-2-8x22b" \| "gryphe/mythomax-l2-13b" \| "meta-llama/llama-4-scout" \| "meta-llama/llama-4-maverick" \| "x-ai/grok-4" \| "x-ai/grok-4-fast" \| "x-ai/grok-4.1-fast" \| "x-ai/grok-code-fast-1" \| "moonshotai/kimi-k2" \| "qwen/qwen3-235b-a22b-thinking-2507" \| "qwen/qwen3-coder" \| "Llama-4-Scout-17B-16E-Instruct-FP8" \| "Llama-4-Maverick-17B-128E-Instruct-FP8" \| "Llama-3.3-8B-Instruct" \| "Llama-3.3-70B-Instruct" \| "v0-1.5-md" \| "v0-1.5-lg" \| "v0-1.0-md" | No |
+| model | The language model to use for the conversation. | "o3-mini" \| "o3-2025-04-16" \| "o1" \| "o1-mini" \| "gpt-5.2-2025-12-11" \| "gpt-5.1-2025-11-13" \| "gpt-5-2025-08-07" \| "gpt-5-mini-2025-08-07" \| "gpt-5-nano-2025-08-07" \| "gpt-5-chat-latest" \| "gpt-4.1-2025-04-14" \| "gpt-4.1-mini-2025-04-14" \| "gpt-4o-mini" \| "gpt-4o" \| "gpt-4-turbo" \| "gpt-3.5-turbo" \| "claude-opus-4-1-20250805" \| "claude-opus-4-20250514" \| "claude-sonnet-4-20250514" \| "claude-opus-4-5-20251101" \| "claude-sonnet-4-5-20250929" \| "claude-haiku-4-5-20251001" \| "claude-3-haiku-20240307" \| "Qwen/Qwen2.5-72B-Instruct-Turbo" \| "nvidia/llama-3.1-nemotron-70b-instruct" \| "meta-llama/Llama-3.3-70B-Instruct-Turbo" \| "meta-llama/Meta-Llama-3.1-70B-Instruct-Turbo" \| "meta-llama/Llama-3.2-3B-Instruct-Turbo" \| "llama-3.3-70b-versatile" \| "llama-3.1-8b-instant" \| "llama3.3" \| "llama3.2" \| "llama3" \| "llama3.1:405b" \| "dolphin-mistral:latest" \| "openai/gpt-oss-120b" \| "openai/gpt-oss-20b" \| "google/gemini-2.5-pro-preview-03-25" \| "google/gemini-3-pro-preview" \| "google/gemini-2.5-flash" \| "google/gemini-2.0-flash-001" \| "google/gemini-2.5-flash-lite-preview-06-17" \| "google/gemini-2.0-flash-lite-001" \| "mistralai/mistral-nemo" \| "cohere/command-r-08-2024" \| "cohere/command-r-plus-08-2024" \| "deepseek/deepseek-chat" \| "deepseek/deepseek-r1-0528" \| "perplexity/sonar" \| "perplexity/sonar-pro" \| "perplexity/sonar-deep-research" \| "nousresearch/hermes-3-llama-3.1-405b" \| "nousresearch/hermes-3-llama-3.1-70b" \| "amazon/nova-lite-v1" \| "amazon/nova-micro-v1" \| "amazon/nova-pro-v1" \| "microsoft/wizardlm-2-8x22b" \| "gryphe/mythomax-l2-13b" \| "meta-llama/llama-4-scout" \| "meta-llama/llama-4-maverick" \| "x-ai/grok-4" \| "x-ai/grok-4-fast" \| "x-ai/grok-4.1-fast" \| "x-ai/grok-code-fast-1" \| "moonshotai/kimi-k2" \| "qwen/qwen3-235b-a22b-thinking-2507" \| "qwen/qwen3-coder" \| "Llama-4-Scout-17B-16E-Instruct-FP8" \| "Llama-4-Maverick-17B-128E-Instruct-FP8" \| "Llama-3.3-8B-Instruct" \| "Llama-3.3-70B-Instruct" \| "v0-1.5-md" \| "v0-1.5-lg" \| "v0-1.0-md" | No |
 | max_tokens | The maximum number of tokens to generate in the chat completion. | int | No |
 | ollama_host | Ollama host for local  models | str | No |
 
@@ -257,7 +257,7 @@ The block formulates a prompt based on the given focus or source data, sends it
 |-------|-------------|------|----------|
 | focus | The focus of the list to generate. | str | No |
 | source_data | The data to generate the list from. | str | No |
-| model | The language model to use for generating the list. | "o3-mini" \| "o3-2025-04-16" \| "o1" \| "o1-mini" \| "gpt-5.2-2025-12-11" \| "gpt-5.1-2025-11-13" \| "gpt-5-2025-08-07" \| "gpt-5-mini-2025-08-07" \| "gpt-5-nano-2025-08-07" \| "gpt-5-chat-latest" \| "gpt-4.1-2025-04-14" \| "gpt-4.1-mini-2025-04-14" \| "gpt-4o-mini" \| "gpt-4o" \| "gpt-4-turbo" \| "gpt-3.5-turbo" \| "claude-opus-4-1-20250805" \| "claude-opus-4-20250514" \| "claude-sonnet-4-20250514" \| "claude-opus-4-5-20251101" \| "claude-sonnet-4-5-20250929" \| "claude-haiku-4-5-20251001" \| "claude-3-7-sonnet-20250219" \| "claude-3-haiku-20240307" \| "Qwen/Qwen2.5-72B-Instruct-Turbo" \| "nvidia/llama-3.1-nemotron-70b-instruct" \| "meta-llama/Llama-3.3-70B-Instruct-Turbo" \| "meta-llama/Meta-Llama-3.1-70B-Instruct-Turbo" \| "meta-llama/Llama-3.2-3B-Instruct-Turbo" \| "llama-3.3-70b-versatile" \| "llama-3.1-8b-instant" \| "llama3.3" \| "llama3.2" \| "llama3" \| "llama3.1:405b" \| "dolphin-mistral:latest" \| "openai/gpt-oss-120b" \| "openai/gpt-oss-20b" \| "google/gemini-2.5-pro-preview-03-25" \| "google/gemini-3-pro-preview" \| "google/gemini-2.5-flash" \| "google/gemini-2.0-flash-001" \| "google/gemini-2.5-flash-lite-preview-06-17" \| "google/gemini-2.0-flash-lite-001" \| "mistralai/mistral-nemo" \| "cohere/command-r-08-2024" \| "cohere/command-r-plus-08-2024" \| "deepseek/deepseek-chat" \| "deepseek/deepseek-r1-0528" \| "perplexity/sonar" \| "perplexity/sonar-pro" \| "perplexity/sonar-deep-research" \| "nousresearch/hermes-3-llama-3.1-405b" \| "nousresearch/hermes-3-llama-3.1-70b" \| "amazon/nova-lite-v1" \| "amazon/nova-micro-v1" \| "amazon/nova-pro-v1" \| "microsoft/wizardlm-2-8x22b" \| "gryphe/mythomax-l2-13b" \| "meta-llama/llama-4-scout" \| "meta-llama/llama-4-maverick" \| "x-ai/grok-4" \| "x-ai/grok-4-fast" \| "x-ai/grok-4.1-fast" \| "x-ai/grok-code-fast-1" \| "moonshotai/kimi-k2" \| "qwen/qwen3-235b-a22b-thinking-2507" \| "qwen/qwen3-coder" \| "Llama-4-Scout-17B-16E-Instruct-FP8" \| "Llama-4-Maverick-17B-128E-Instruct-FP8" \| "Llama-3.3-8B-Instruct" \| "Llama-3.3-70B-Instruct" \| "v0-1.5-md" \| "v0-1.5-lg" \| "v0-1.0-md" | No |
+| model | The language model to use for generating the list. | "o3-mini" \| "o3-2025-04-16" \| "o1" \| "o1-mini" \| "gpt-5.2-2025-12-11" \| "gpt-5.1-2025-11-13" \| "gpt-5-2025-08-07" \| "gpt-5-mini-2025-08-07" \| "gpt-5-nano-2025-08-07" \| "gpt-5-chat-latest" \| "gpt-4.1-2025-04-14" \| "gpt-4.1-mini-2025-04-14" \| "gpt-4o-mini" \| "gpt-4o" \| "gpt-4-turbo" \| "gpt-3.5-turbo" \| "claude-opus-4-1-20250805" \| "claude-opus-4-20250514" \| "claude-sonnet-4-20250514" \| "claude-opus-4-5-20251101" \| "claude-sonnet-4-5-20250929" \| "claude-haiku-4-5-20251001" \| "claude-3-haiku-20240307" \| "Qwen/Qwen2.5-72B-Instruct-Turbo" \| "nvidia/llama-3.1-nemotron-70b-instruct" \| "meta-llama/Llama-3.3-70B-Instruct-Turbo" \| "meta-llama/Meta-Llama-3.1-70B-Instruct-Turbo" \| "meta-llama/Llama-3.2-3B-Instruct-Turbo" \| "llama-3.3-70b-versatile" \| "llama-3.1-8b-instant" \| "llama3.3" \| "llama3.2" \| "llama3" \| "llama3.1:405b" \| "dolphin-mistral:latest" \| "openai/gpt-oss-120b" \| "openai/gpt-oss-20b" \| "google/gemini-2.5-pro-preview-03-25" \| "google/gemini-3-pro-preview" \| "google/gemini-2.5-flash" \| "google/gemini-2.0-flash-001" \| "google/gemini-2.5-flash-lite-preview-06-17" \| "google/gemini-2.0-flash-lite-001" \| "mistralai/mistral-nemo" \| "cohere/command-r-08-2024" \| "cohere/command-r-plus-08-2024" \| "deepseek/deepseek-chat" \| "deepseek/deepseek-r1-0528" \| "perplexity/sonar" \| "perplexity/sonar-pro" \| "perplexity/sonar-deep-research" \| "nousresearch/hermes-3-llama-3.1-405b" \| "nousresearch/hermes-3-llama-3.1-70b" \| "amazon/nova-lite-v1" \| "amazon/nova-micro-v1" \| "amazon/nova-pro-v1" \| "microsoft/wizardlm-2-8x22b" \| "gryphe/mythomax-l2-13b" \| "meta-llama/llama-4-scout" \| "meta-llama/llama-4-maverick" \| "x-ai/grok-4" \| "x-ai/grok-4-fast" \| "x-ai/grok-4.1-fast" \| "x-ai/grok-code-fast-1" \| "moonshotai/kimi-k2" \| "qwen/qwen3-235b-a22b-thinking-2507" \| "qwen/qwen3-coder" \| "Llama-4-Scout-17B-16E-Instruct-FP8" \| "Llama-4-Maverick-17B-128E-Instruct-FP8" \| "Llama-3.3-8B-Instruct" \| "Llama-3.3-70B-Instruct" \| "v0-1.5-md" \| "v0-1.5-lg" \| "v0-1.0-md" | No |
 | max_retries | Maximum number of retries for generating a valid list. | int | No |
 | force_json_output | Whether to force the LLM to produce a JSON-only response. This can increase the block's reliability, but may also reduce the quality of the response because it prohibits the LLM from reasoning before providing its JSON response. | bool | No |
 | max_tokens | The maximum number of tokens to generate in the chat completion. | int | No |
@@ -424,7 +424,7 @@ The block sends the input prompt to a chosen LLM, along with any system prompts
 | prompt | The prompt to send to the language model. | str | Yes |
 | expected_format | Expected format of the response. If provided, the response will be validated against this format. The keys should be the expected fields in the response, and the values should be the description of the field. | Dict[str, str] | Yes |
 | list_result | Whether the response should be a list of objects in the expected format. | bool | No |
-| model | The language model to use for answering the prompt. | "o3-mini" \| "o3-2025-04-16" \| "o1" \| "o1-mini" \| "gpt-5.2-2025-12-11" \| "gpt-5.1-2025-11-13" \| "gpt-5-2025-08-07" \| "gpt-5-mini-2025-08-07" \| "gpt-5-nano-2025-08-07" \| "gpt-5-chat-latest" \| "gpt-4.1-2025-04-14" \| "gpt-4.1-mini-2025-04-14" \| "gpt-4o-mini" \| "gpt-4o" \| "gpt-4-turbo" \| "gpt-3.5-turbo" \| "claude-opus-4-1-20250805" \| "claude-opus-4-20250514" \| "claude-sonnet-4-20250514" \| "claude-opus-4-5-20251101" \| "claude-sonnet-4-5-20250929" \| "claude-haiku-4-5-20251001" \| "claude-3-7-sonnet-20250219" \| "claude-3-haiku-20240307" \| "Qwen/Qwen2.5-72B-Instruct-Turbo" \| "nvidia/llama-3.1-nemotron-70b-instruct" \| "meta-llama/Llama-3.3-70B-Instruct-Turbo" \| "meta-llama/Meta-Llama-3.1-70B-Instruct-Turbo" \| "meta-llama/Llama-3.2-3B-Instruct-Turbo" \| "llama-3.3-70b-versatile" \| "llama-3.1-8b-instant" \| "llama3.3" \| "llama3.2" \| "llama3" \| "llama3.1:405b" \| "dolphin-mistral:latest" \| "openai/gpt-oss-120b" \| "openai/gpt-oss-20b" \| "google/gemini-2.5-pro-preview-03-25" \| "google/gemini-3-pro-preview" \| "google/gemini-2.5-flash" \| "google/gemini-2.0-flash-001" \| "google/gemini-2.5-flash-lite-preview-06-17" \| "google/gemini-2.0-flash-lite-001" \| "mistralai/mistral-nemo" \| "cohere/command-r-08-2024" \| "cohere/command-r-plus-08-2024" \| "deepseek/deepseek-chat" \| "deepseek/deepseek-r1-0528" \| "perplexity/sonar" \| "perplexity/sonar-pro" \| "perplexity/sonar-deep-research" \| "nousresearch/hermes-3-llama-3.1-405b" \| "nousresearch/hermes-3-llama-3.1-70b" \| "amazon/nova-lite-v1" \| "amazon/nova-micro-v1" \| "amazon/nova-pro-v1" \| "microsoft/wizardlm-2-8x22b" \| "gryphe/mythomax-l2-13b" \| "meta-llama/llama-4-scout" \| "meta-llama/llama-4-maverick" \| "x-ai/grok-4" \| "x-ai/grok-4-fast" \| "x-ai/grok-4.1-fast" \| "x-ai/grok-code-fast-1" \| "moonshotai/kimi-k2" \| "qwen/qwen3-235b-a22b-thinking-2507" \| "qwen/qwen3-coder" \| "Llama-4-Scout-17B-16E-Instruct-FP8" \| "Llama-4-Maverick-17B-128E-Instruct-FP8" \| "Llama-3.3-8B-Instruct" \| "Llama-3.3-70B-Instruct" \| "v0-1.5-md" \| "v0-1.5-lg" \| "v0-1.0-md" | No |
+| model | The language model to use for answering the prompt. | "o3-mini" \| "o3-2025-04-16" \| "o1" \| "o1-mini" \| "gpt-5.2-2025-12-11" \| "gpt-5.1-2025-11-13" \| "gpt-5-2025-08-07" \| "gpt-5-mini-2025-08-07" \| "gpt-5-nano-2025-08-07" \| "gpt-5-chat-latest" \| "gpt-4.1-2025-04-14" \| "gpt-4.1-mini-2025-04-14" \| "gpt-4o-mini" \| "gpt-4o" \| "gpt-4-turbo" \| "gpt-3.5-turbo" \| "claude-opus-4-1-20250805" \| "claude-opus-4-20250514" \| "claude-sonnet-4-20250514" \| "claude-opus-4-5-20251101" \| "claude-sonnet-4-5-20250929" \| "claude-haiku-4-5-20251001" \| "claude-3-haiku-20240307" \| "Qwen/Qwen2.5-72B-Instruct-Turbo" \| "nvidia/llama-3.1-nemotron-70b-instruct" \| "meta-llama/Llama-3.3-70B-Instruct-Turbo" \| "meta-llama/Meta-Llama-3.1-70B-Instruct-Turbo" \| "meta-llama/Llama-3.2-3B-Instruct-Turbo" \| "llama-3.3-70b-versatile" \| "llama-3.1-8b-instant" \| "llama3.3" \| "llama3.2" \| "llama3" \| "llama3.1:405b" \| "dolphin-mistral:latest" \| "openai/gpt-oss-120b" \| "openai/gpt-oss-20b" \| "google/gemini-2.5-pro-preview-03-25" \| "google/gemini-3-pro-preview" \| "google/gemini-2.5-flash" \| "google/gemini-2.0-flash-001" \| "google/gemini-2.5-flash-lite-preview-06-17" \| "google/gemini-2.0-flash-lite-001" \| "mistralai/mistral-nemo" \| "cohere/command-r-08-2024" \| "cohere/command-r-plus-08-2024" \| "deepseek/deepseek-chat" \| "deepseek/deepseek-r1-0528" \| "perplexity/sonar" \| "perplexity/sonar-pro" \| "perplexity/sonar-deep-research" \| "nousresearch/hermes-3-llama-3.1-405b" \| "nousresearch/hermes-3-llama-3.1-70b" \| "amazon/nova-lite-v1" \| "amazon/nova-micro-v1" \| "amazon/nova-pro-v1" \| "microsoft/wizardlm-2-8x22b" \| "gryphe/mythomax-l2-13b" \| "meta-llama/llama-4-scout" \| "meta-llama/llama-4-maverick" \| "x-ai/grok-4" \| "x-ai/grok-4-fast" \| "x-ai/grok-4.1-fast" \| "x-ai/grok-code-fast-1" \| "moonshotai/kimi-k2" \| "qwen/qwen3-235b-a22b-thinking-2507" \| "qwen/qwen3-coder" \| "Llama-4-Scout-17B-16E-Instruct-FP8" \| "Llama-4-Maverick-17B-128E-Instruct-FP8" \| "Llama-3.3-8B-Instruct" \| "Llama-3.3-70B-Instruct" \| "v0-1.5-md" \| "v0-1.5-lg" \| "v0-1.0-md" | No |
 | force_json_output | Whether to force the LLM to produce a JSON-only response. This can increase the block's reliability, but may also reduce the quality of the response because it prohibits the LLM from reasoning before providing its JSON response. | bool | No |
 | sys_prompt | The system prompt to provide additional context to the model. | str | No |
 | conversation_history | The conversation history to provide context for the prompt. | List[Dict[str, Any]] | No |
@@ -464,7 +464,7 @@ The block sends the input prompt to a chosen LLM, processes the response, and re
 | Input | Description | Type | Required |
 |-------|-------------|------|----------|
 | prompt | The prompt to send to the language model. You can use any of the {keys} from Prompt Values to fill in the prompt with values from the prompt values dictionary by putting them in curly braces. | str | Yes |
-| model | The language model to use for answering the prompt. | "o3-mini" \| "o3-2025-04-16" \| "o1" \| "o1-mini" \| "gpt-5.2-2025-12-11" \| "gpt-5.1-2025-11-13" \| "gpt-5-2025-08-07" \| "gpt-5-mini-2025-08-07" \| "gpt-5-nano-2025-08-07" \| "gpt-5-chat-latest" \| "gpt-4.1-2025-04-14" \| "gpt-4.1-mini-2025-04-14" \| "gpt-4o-mini" \| "gpt-4o" \| "gpt-4-turbo" \| "gpt-3.5-turbo" \| "claude-opus-4-1-20250805" \| "claude-opus-4-20250514" \| "claude-sonnet-4-20250514" \| "claude-opus-4-5-20251101" \| "claude-sonnet-4-5-20250929" \| "claude-haiku-4-5-20251001" \| "claude-3-7-sonnet-20250219" \| "claude-3-haiku-20240307" \| "Qwen/Qwen2.5-72B-Instruct-Turbo" \| "nvidia/llama-3.1-nemotron-70b-instruct" \| "meta-llama/Llama-3.3-70B-Instruct-Turbo" \| "meta-llama/Meta-Llama-3.1-70B-Instruct-Turbo" \| "meta-llama/Llama-3.2-3B-Instruct-Turbo" \| "llama-3.3-70b-versatile" \| "llama-3.1-8b-instant" \| "llama3.3" \| "llama3.2" \| "llama3" \| "llama3.1:405b" \| "dolphin-mistral:latest" \| "openai/gpt-oss-120b" \| "openai/gpt-oss-20b" \| "google/gemini-2.5-pro-preview-03-25" \| "google/gemini-3-pro-preview" \| "google/gemini-2.5-flash" \| "google/gemini-2.0-flash-001" \| "google/gemini-2.5-flash-lite-preview-06-17" \| "google/gemini-2.0-flash-lite-001" \| "mistralai/mistral-nemo" \| "cohere/command-r-08-2024" \| "cohere/command-r-plus-08-2024" \| "deepseek/deepseek-chat" \| "deepseek/deepseek-r1-0528" \| "perplexity/sonar" \| "perplexity/sonar-pro" \| "perplexity/sonar-deep-research" \| "nousresearch/hermes-3-llama-3.1-405b" \| "nousresearch/hermes-3-llama-3.1-70b" \| "amazon/nova-lite-v1" \| "amazon/nova-micro-v1" \| "amazon/nova-pro-v1" \| "microsoft/wizardlm-2-8x22b" \| "gryphe/mythomax-l2-13b" \| "meta-llama/llama-4-scout" \| "meta-llama/llama-4-maverick" \| "x-ai/grok-4" \| "x-ai/grok-4-fast" \| "x-ai/grok-4.1-fast" \| "x-ai/grok-code-fast-1" \| "moonshotai/kimi-k2" \| "qwen/qwen3-235b-a22b-thinking-2507" \| "qwen/qwen3-coder" \| "Llama-4-Scout-17B-16E-Instruct-FP8" \| "Llama-4-Maverick-17B-128E-Instruct-FP8" \| "Llama-3.3-8B-Instruct" \| "Llama-3.3-70B-Instruct" \| "v0-1.5-md" \| "v0-1.5-lg" \| "v0-1.0-md" | No |
+| model | The language model to use for answering the prompt. | "o3-mini" \| "o3-2025-04-16" \| "o1" \| "o1-mini" \| "gpt-5.2-2025-12-11" \| "gpt-5.1-2025-11-13" \| "gpt-5-2025-08-07" \| "gpt-5-mini-2025-08-07" \| "gpt-5-nano-2025-08-07" \| "gpt-5-chat-latest" \| "gpt-4.1-2025-04-14" \| "gpt-4.1-mini-2025-04-14" \| "gpt-4o-mini" \| "gpt-4o" \| "gpt-4-turbo" \| "gpt-3.5-turbo" \| "claude-opus-4-1-20250805" \| "claude-opus-4-20250514" \| "claude-sonnet-4-20250514" \| "claude-opus-4-5-20251101" \| "claude-sonnet-4-5-20250929" \| "claude-haiku-4-5-20251001" \| "claude-3-haiku-20240307" \| "Qwen/Qwen2.5-72B-Instruct-Turbo" \| "nvidia/llama-3.1-nemotron-70b-instruct" \| "meta-llama/Llama-3.3-70B-Instruct-Turbo" \| "meta-llama/Meta-Llama-3.1-70B-Instruct-Turbo" \| "meta-llama/Llama-3.2-3B-Instruct-Turbo" \| "llama-3.3-70b-versatile" \| "llama-3.1-8b-instant" \| "llama3.3" \| "llama3.2" \| "llama3" \| "llama3.1:405b" \| "dolphin-mistral:latest" \| "openai/gpt-oss-120b" \| "openai/gpt-oss-20b" \| "google/gemini-2.5-pro-preview-03-25" \| "google/gemini-3-pro-preview" \| "google/gemini-2.5-flash" \| "google/gemini-2.0-flash-001" \| "google/gemini-2.5-flash-lite-preview-06-17" \| "google/gemini-2.0-flash-lite-001" \| "mistralai/mistral-nemo" \| "cohere/command-r-08-2024" \| "cohere/command-r-plus-08-2024" \| "deepseek/deepseek-chat" \| "deepseek/deepseek-r1-0528" \| "perplexity/sonar" \| "perplexity/sonar-pro" \| "perplexity/sonar-deep-research" \| "nousresearch/hermes-3-llama-3.1-405b" \| "nousresearch/hermes-3-llama-3.1-70b" \| "amazon/nova-lite-v1" \| "amazon/nova-micro-v1" \| "amazon/nova-pro-v1" \| "microsoft/wizardlm-2-8x22b" \| "gryphe/mythomax-l2-13b" \| "meta-llama/llama-4-scout" \| "meta-llama/llama-4-maverick" \| "x-ai/grok-4" \| "x-ai/grok-4-fast" \| "x-ai/grok-4.1-fast" \| "x-ai/grok-code-fast-1" \| "moonshotai/kimi-k2" \| "qwen/qwen3-235b-a22b-thinking-2507" \| "qwen/qwen3-coder" \| "Llama-4-Scout-17B-16E-Instruct-FP8" \| "Llama-4-Maverick-17B-128E-Instruct-FP8" \| "Llama-3.3-8B-Instruct" \| "Llama-3.3-70B-Instruct" \| "v0-1.5-md" \| "v0-1.5-lg" \| "v0-1.0-md" | No |
 | sys_prompt | The system prompt to provide additional context to the model. | str | No |
 | retry | Number of times to retry the LLM call if the response does not match the expected format. | int | No |
 | prompt_values | Values used to fill in the prompt. The values can be used in the prompt by putting them in a double curly braces, e.g. {{variable_name}}. | Dict[str, str] | No |
@@ -501,7 +501,7 @@ The block splits the input text into smaller chunks, sends each chunk to an LLM
 | Input | Description | Type | Required |
 |-------|-------------|------|----------|
 | text | The text to summarize. | str | Yes |
-| model | The language model to use for summarizing the text. | "o3-mini" \| "o3-2025-04-16" \| "o1" \| "o1-mini" \| "gpt-5.2-2025-12-11" \| "gpt-5.1-2025-11-13" \| "gpt-5-2025-08-07" \| "gpt-5-mini-2025-08-07" \| "gpt-5-nano-2025-08-07" \| "gpt-5-chat-latest" \| "gpt-4.1-2025-04-14" \| "gpt-4.1-mini-2025-04-14" \| "gpt-4o-mini" \| "gpt-4o" \| "gpt-4-turbo" \| "gpt-3.5-turbo" \| "claude-opus-4-1-20250805" \| "claude-opus-4-20250514" \| "claude-sonnet-4-20250514" \| "claude-opus-4-5-20251101" \| "claude-sonnet-4-5-20250929" \| "claude-haiku-4-5-20251001" \| "claude-3-7-sonnet-20250219" \| "claude-3-haiku-20240307" \| "Qwen/Qwen2.5-72B-Instruct-Turbo" \| "nvidia/llama-3.1-nemotron-70b-instruct" \| "meta-llama/Llama-3.3-70B-Instruct-Turbo" \| "meta-llama/Meta-Llama-3.1-70B-Instruct-Turbo" \| "meta-llama/Llama-3.2-3B-Instruct-Turbo" \| "llama-3.3-70b-versatile" \| "llama-3.1-8b-instant" \| "llama3.3" \| "llama3.2" \| "llama3" \| "llama3.1:405b" \| "dolphin-mistral:latest" \| "openai/gpt-oss-120b" \| "openai/gpt-oss-20b" \| "google/gemini-2.5-pro-preview-03-25" \| "google/gemini-3-pro-preview" \| "google/gemini-2.5-flash" \| "google/gemini-2.0-flash-001" \| "google/gemini-2.5-flash-lite-preview-06-17" \| "google/gemini-2.0-flash-lite-001" \| "mistralai/mistral-nemo" \| "cohere/command-r-08-2024" \| "cohere/command-r-plus-08-2024" \| "deepseek/deepseek-chat" \| "deepseek/deepseek-r1-0528" \| "perplexity/sonar" \| "perplexity/sonar-pro" \| "perplexity/sonar-deep-research" \| "nousresearch/hermes-3-llama-3.1-405b" \| "nousresearch/hermes-3-llama-3.1-70b" \| "amazon/nova-lite-v1" \| "amazon/nova-micro-v1" \| "amazon/nova-pro-v1" \| "microsoft/wizardlm-2-8x22b" \| "gryphe/mythomax-l2-13b" \| "meta-llama/llama-4-scout" \| "meta-llama/llama-4-maverick" \| "x-ai/grok-4" \| "x-ai/grok-4-fast" \| "x-ai/grok-4.1-fast" \| "x-ai/grok-code-fast-1" \| "moonshotai/kimi-k2" \| "qwen/qwen3-235b-a22b-thinking-2507" \| "qwen/qwen3-coder" \| "Llama-4-Scout-17B-16E-Instruct-FP8" \| "Llama-4-Maverick-17B-128E-Instruct-FP8" \| "Llama-3.3-8B-Instruct" \| "Llama-3.3-70B-Instruct" \| "v0-1.5-md" \| "v0-1.5-lg" \| "v0-1.0-md" | No |
+| model | The language model to use for summarizing the text. | "o3-mini" \| "o3-2025-04-16" \| "o1" \| "o1-mini" \| "gpt-5.2-2025-12-11" \| "gpt-5.1-2025-11-13" \| "gpt-5-2025-08-07" \| "gpt-5-mini-2025-08-07" \| "gpt-5-nano-2025-08-07" \| "gpt-5-chat-latest" \| "gpt-4.1-2025-04-14" \| "gpt-4.1-mini-2025-04-14" \| "gpt-4o-mini" \| "gpt-4o" \| "gpt-4-turbo" \| "gpt-3.5-turbo" \| "claude-opus-4-1-20250805" \| "claude-opus-4-20250514" \| "claude-sonnet-4-20250514" \| "claude-opus-4-5-20251101" \| "claude-sonnet-4-5-20250929" \| "claude-haiku-4-5-20251001" \| "claude-3-haiku-20240307" \| "Qwen/Qwen2.5-72B-Instruct-Turbo" \| "nvidia/llama-3.1-nemotron-70b-instruct" \| "meta-llama/Llama-3.3-70B-Instruct-Turbo" \| "meta-llama/Meta-Llama-3.1-70B-Instruct-Turbo" \| "meta-llama/Llama-3.2-3B-Instruct-Turbo" \| "llama-3.3-70b-versatile" \| "llama-3.1-8b-instant" \| "llama3.3" \| "llama3.2" \| "llama3" \| "llama3.1:405b" \| "dolphin-mistral:latest" \| "openai/gpt-oss-120b" \| "openai/gpt-oss-20b" \| "google/gemini-2.5-pro-preview-03-25" \| "google/gemini-3-pro-preview" \| "google/gemini-2.5-flash" \| "google/gemini-2.0-flash-001" \| "google/gemini-2.5-flash-lite-preview-06-17" \| "google/gemini-2.0-flash-lite-001" \| "mistralai/mistral-nemo" \| "cohere/command-r-08-2024" \| "cohere/command-r-plus-08-2024" \| "deepseek/deepseek-chat" \| "deepseek/deepseek-r1-0528" \| "perplexity/sonar" \| "perplexity/sonar-pro" \| "perplexity/sonar-deep-research" \| "nousresearch/hermes-3-llama-3.1-405b" \| "nousresearch/hermes-3-llama-3.1-70b" \| "amazon/nova-lite-v1" \| "amazon/nova-micro-v1" \| "amazon/nova-pro-v1" \| "microsoft/wizardlm-2-8x22b" \| "gryphe/mythomax-l2-13b" \| "meta-llama/llama-4-scout" \| "meta-llama/llama-4-maverick" \| "x-ai/grok-4" \| "x-ai/grok-4-fast" \| "x-ai/grok-4.1-fast" \| "x-ai/grok-code-fast-1" \| "moonshotai/kimi-k2" \| "qwen/qwen3-235b-a22b-thinking-2507" \| "qwen/qwen3-coder" \| "Llama-4-Scout-17B-16E-Instruct-FP8" \| "Llama-4-Maverick-17B-128E-Instruct-FP8" \| "Llama-3.3-8B-Instruct" \| "Llama-3.3-70B-Instruct" \| "v0-1.5-md" \| "v0-1.5-lg" \| "v0-1.0-md" | No |
 | focus | The topic to focus on in the summary | str | No |
 | style | The style of the summary to generate. | "concise" \| "detailed" \| "bullet points" \| "numbered list" | No |
 | max_tokens | The maximum number of tokens to generate in the chat completion. | int | No |
@@ -763,7 +763,7 @@ Configure agent_mode_max_iterations to control loop behavior: 0 for single decis
 | Input | Description | Type | Required |
 |-------|-------------|------|----------|
 | prompt | The prompt to send to the language model. | str | Yes |
-| model | The language model to use for answering the prompt. | "o3-mini" \| "o3-2025-04-16" \| "o1" \| "o1-mini" \| "gpt-5.2-2025-12-11" \| "gpt-5.1-2025-11-13" \| "gpt-5-2025-08-07" \| "gpt-5-mini-2025-08-07" \| "gpt-5-nano-2025-08-07" \| "gpt-5-chat-latest" \| "gpt-4.1-2025-04-14" \| "gpt-4.1-mini-2025-04-14" \| "gpt-4o-mini" \| "gpt-4o" \| "gpt-4-turbo" \| "gpt-3.5-turbo" \| "claude-opus-4-1-20250805" \| "claude-opus-4-20250514" \| "claude-sonnet-4-20250514" \| "claude-opus-4-5-20251101" \| "claude-sonnet-4-5-20250929" \| "claude-haiku-4-5-20251001" \| "claude-3-7-sonnet-20250219" \| "claude-3-haiku-20240307" \| "Qwen/Qwen2.5-72B-Instruct-Turbo" \| "nvidia/llama-3.1-nemotron-70b-instruct" \| "meta-llama/Llama-3.3-70B-Instruct-Turbo" \| "meta-llama/Meta-Llama-3.1-70B-Instruct-Turbo" \| "meta-llama/Llama-3.2-3B-Instruct-Turbo" \| "llama-3.3-70b-versatile" \| "llama-3.1-8b-instant" \| "llama3.3" \| "llama3.2" \| "llama3" \| "llama3.1:405b" \| "dolphin-mistral:latest" \| "openai/gpt-oss-120b" \| "openai/gpt-oss-20b" \| "google/gemini-2.5-pro-preview-03-25" \| "google/gemini-3-pro-preview" \| "google/gemini-2.5-flash" \| "google/gemini-2.0-flash-001" \| "google/gemini-2.5-flash-lite-preview-06-17" \| "google/gemini-2.0-flash-lite-001" \| "mistralai/mistral-nemo" \| "cohere/command-r-08-2024" \| "cohere/command-r-plus-08-2024" \| "deepseek/deepseek-chat" \| "deepseek/deepseek-r1-0528" \| "perplexity/sonar" \| "perplexity/sonar-pro" \| "perplexity/sonar-deep-research" \| "nousresearch/hermes-3-llama-3.1-405b" \| "nousresearch/hermes-3-llama-3.1-70b" \| "amazon/nova-lite-v1" \| "amazon/nova-micro-v1" \| "amazon/nova-pro-v1" \| "microsoft/wizardlm-2-8x22b" \| "gryphe/mythomax-l2-13b" \| "meta-llama/llama-4-scout" \| "meta-llama/llama-4-maverick" \| "x-ai/grok-4" \| "x-ai/grok-4-fast" \| "x-ai/grok-4.1-fast" \| "x-ai/grok-code-fast-1" \| "moonshotai/kimi-k2" \| "qwen/qwen3-235b-a22b-thinking-2507" \| "qwen/qwen3-coder" \| "Llama-4-Scout-17B-16E-Instruct-FP8" \| "Llama-4-Maverick-17B-128E-Instruct-FP8" \| "Llama-3.3-8B-Instruct" \| "Llama-3.3-70B-Instruct" \| "v0-1.5-md" \| "v0-1.5-lg" \| "v0-1.0-md" | No |
+| model | The language model to use for answering the prompt. | "o3-mini" \| "o3-2025-04-16" \| "o1" \| "o1-mini" \| "gpt-5.2-2025-12-11" \| "gpt-5.1-2025-11-13" \| "gpt-5-2025-08-07" \| "gpt-5-mini-2025-08-07" \| "gpt-5-nano-2025-08-07" \| "gpt-5-chat-latest" \| "gpt-4.1-2025-04-14" \| "gpt-4.1-mini-2025-04-14" \| "gpt-4o-mini" \| "gpt-4o" \| "gpt-4-turbo" \| "gpt-3.5-turbo" \| "claude-opus-4-1-20250805" \| "claude-opus-4-20250514" \| "claude-sonnet-4-20250514" \| "claude-opus-4-5-20251101" \| "claude-sonnet-4-5-20250929" \| "claude-haiku-4-5-20251001" \| "claude-3-haiku-20240307" \| "Qwen/Qwen2.5-72B-Instruct-Turbo" \| "nvidia/llama-3.1-nemotron-70b-instruct" \| "meta-llama/Llama-3.3-70B-Instruct-Turbo" \| "meta-llama/Meta-Llama-3.1-70B-Instruct-Turbo" \| "meta-llama/Llama-3.2-3B-Instruct-Turbo" \| "llama-3.3-70b-versatile" \| "llama-3.1-8b-instant" \| "llama3.3" \| "llama3.2" \| "llama3" \| "llama3.1:405b" \| "dolphin-mistral:latest" \| "openai/gpt-oss-120b" \| "openai/gpt-oss-20b" \| "google/gemini-2.5-pro-preview-03-25" \| "google/gemini-3-pro-preview" \| "google/gemini-2.5-flash" \| "google/gemini-2.0-flash-001" \| "google/gemini-2.5-flash-lite-preview-06-17" \| "google/gemini-2.0-flash-lite-001" \| "mistralai/mistral-nemo" \| "cohere/command-r-08-2024" \| "cohere/command-r-plus-08-2024" \| "deepseek/deepseek-chat" \| "deepseek/deepseek-r1-0528" \| "perplexity/sonar" \| "perplexity/sonar-pro" \| "perplexity/sonar-deep-research" \| "nousresearch/hermes-3-llama-3.1-405b" \| "nousresearch/hermes-3-llama-3.1-70b" \| "amazon/nova-lite-v1" \| "amazon/nova-micro-v1" \| "amazon/nova-pro-v1" \| "microsoft/wizardlm-2-8x22b" \| "gryphe/mythomax-l2-13b" \| "meta-llama/llama-4-scout" \| "meta-llama/llama-4-maverick" \| "x-ai/grok-4" \| "x-ai/grok-4-fast" \| "x-ai/grok-4.1-fast" \| "x-ai/grok-code-fast-1" \| "moonshotai/kimi-k2" \| "qwen/qwen3-235b-a22b-thinking-2507" \| "qwen/qwen3-coder" \| "Llama-4-Scout-17B-16E-Instruct-FP8" \| "Llama-4-Maverick-17B-128E-Instruct-FP8" \| "Llama-3.3-8B-Instruct" \| "Llama-3.3-70B-Instruct" \| "v0-1.5-md" \| "v0-1.5-lg" \| "v0-1.0-md" | No |
 | multiple_tool_calls | Whether to allow multiple tool calls in a single response. | bool | No |
 | sys_prompt | The system prompt to provide additional context to the model. | str | No |
 | conversation_history | The conversation history to provide context for the prompt. | List[Dict[str, Any]] | No |

docs/integrations/block-integrations/stagehand/blocks.md

@@ -20,7 +20,7 @@ Configure timeouts for DOM settlement and page loading. Variables can be passed
 | Input | Description | Type | Required |
 |-------|-------------|------|----------|
 | browserbase_project_id | Browserbase project ID (required if using Browserbase) | str | Yes |
-| model | LLM to use for Stagehand (provider is inferred) | "gpt-4.1-2025-04-14" \| "gpt-4.1-mini-2025-04-14" \| "claude-3-7-sonnet-20250219" | No |
+| model | LLM to use for Stagehand (provider is inferred) | "gpt-4.1-2025-04-14" \| "gpt-4.1-mini-2025-04-14" \| "claude-sonnet-4-5-20250929" | No |
 | url | URL to navigate to. | str | Yes |
 | action | Action to perform. Suggested actions are: click, fill, type, press, scroll, select from dropdown. For multi-step actions, add an entry for each step. | List[str] | Yes |
 | variables | Variables to use in the action. Variables contains data you want the action to use. | Dict[str, str] | No |
@@ -65,7 +65,7 @@ Supports searching within iframes and configurable timeouts for dynamic content
 | Input | Description | Type | Required |
 |-------|-------------|------|----------|
 | browserbase_project_id | Browserbase project ID (required if using Browserbase) | str | Yes |
-| model | LLM to use for Stagehand (provider is inferred) | "gpt-4.1-2025-04-14" \| "gpt-4.1-mini-2025-04-14" \| "claude-3-7-sonnet-20250219" | No |
+| model | LLM to use for Stagehand (provider is inferred) | "gpt-4.1-2025-04-14" \| "gpt-4.1-mini-2025-04-14" \| "claude-sonnet-4-5-20250929" | No |
 | url | URL to navigate to. | str | Yes |
 | instruction | Natural language description of elements or actions to discover. | str | Yes |
 | iframes | Whether to search within iframes. If True, Stagehand will search for actions within iframes. | bool | No |
@@ -106,7 +106,7 @@ Use this to explore a page's interactive elements before building automated work
 | Input | Description | Type | Required |
 |-------|-------------|------|----------|
 | browserbase_project_id | Browserbase project ID (required if using Browserbase) | str | Yes |
-| model | LLM to use for Stagehand (provider is inferred) | "gpt-4.1-2025-04-14" \| "gpt-4.1-mini-2025-04-14" \| "claude-3-7-sonnet-20250219" | No |
+| model | LLM to use for Stagehand (provider is inferred) | "gpt-4.1-2025-04-14" \| "gpt-4.1-mini-2025-04-14" \| "claude-sonnet-4-5-20250929" | No |
 | url | URL to navigate to. | str | Yes |
 | instruction | Natural language description of elements or actions to discover. | str | Yes |
 | iframes | Whether to search within iframes. If True, Stagehand will search for actions within iframes. | bool | No |