Add-on: LangGraph Agent

Use this skill when a project needs multi-step agent execution (tool use, checkpoints, and bounded loops) through LangGraph.

Compatibility

• Works with architect-python-uv-fastapi-sqlalchemy and architect-nextjs-bun-app.
• Can be combined with addon-langchain-llm for shared provider setup.
• Can be combined with addon-google-agent-dev-kit when ADK interop is needed.

Inputs

Collect:

• AGENT_SCOPE: tasks the agent is explicitly allowed to perform.
• CHECKPOINT_BACKEND: memory | postgres | redis.
• ALLOW_TOOLS: explicit tool allow-list.
• MAX_STEPS: default 10.
• TIMEOUT_SECONDS: default 45.