langgraph-subgraphs
Pass
Audited by Gen Agent Trust Hub on Feb 17, 2026
Risk Level: SAFE
Full Analysis
- SAFE (SAFE): The skill provides structural templates for LangGraph subgraphs. Analysis shows no malicious patterns, unauthorized file access, or network exfiltration.
- PROMPT_INJECTION (LOW): The skill describes patterns for processing external input (Indirect Prompt Injection surface). 1. Ingestion points: User input enters via 'query' and 'input_text' fields in ParentState and AnalysisState definitions in SKILL.md and references/invoke-pattern.md. 2. Boundary markers: None explicitly defined in the primary templates to isolate user data from instructions. 3. Capability inventory: Uses 'subgraph.invoke()' to execute graph logic in nodes. 4. Sanitization: The 'references/state-mapping.md' file provides examples of input validation using Pydantic, which is a recommended mitigation.
Audit Metadata