langgraph
Pass
Audited by Gen Agent Trust Hub on Jul 19, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: The skill provides legitimate instructions and tools for developing AI agents with the LangGraph framework. All code and scripts align with the stated purpose of scaffolding and documenting agentic workflows.
- [REMOTE_CODE_EXECUTION]: Heuristic analysis flagged 'compile()' calls in 'assets/templates/subgraph-agent.py'. Manual review confirms these are calls to 'langgraph.graph.StateGraph.compile()', which is a standard framework method for finalizing graph structures. It does not perform dynamic execution of arbitrary strings or shell commands.
- [EXTERNAL_DOWNLOADS]: The skill instructs the user to install well-known and reputable Python packages from the LangChain ecosystem, including 'langgraph', 'langchain', 'langchain-openai', and 'langsmith'.
- [COMMAND_EXECUTION]: The scaffolding scripts in the 'scripts/' directory are designed to generate local project files. This behavior is expected for developer tools and is implemented safely using static templates without untrusted data flowing into dangerous sinks.
Audit Metadata