attack-path-stitcher

Pass

Audited by Gen Agent Trust Hub on Jun 13, 2026

Risk Level: SAFEPROMPT_INJECTION
Full Analysis
  • [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted data from external findings which could potentially contain malicious instructions.
  • Ingestion points: Reads data from validated/*.json and findings/*/evidence/raw-source.txt (containing raw evidence from various analyzed assets).
  • Boundary markers: None explicitly defined in the skill instructions to separate processed data from agent instructions during analysis.
  • Capability inventory: The skill executes tools/chain-merger.py to perform graph construction and writes findings to artifacts/attack-paths.json, artifacts/attack-paths.dot, and artifacts/attack-paths.md.
  • Sanitization: No sanitization or validation logic is described for the raw evidence strings extracted via regex patterns before they are incorporated into the output reports.
  • [DATA_EXPOSURE]: The skill is designed to identify and consolidate sensitive information, including credentials (passwords, JWTs, hashes, API keys) and organizational trust relationships. While this is an intended function for an attack path analysis tool, it involves handling and centralizing high-value sensitive data.
Audit Metadata
Risk Level
SAFE
Analyzed
Jun 13, 2026, 11:34 PM
Security Audit — agent-trust-hub — attack-path-stitcher