retro

Pass

Audited by Gen Agent Trust Hub on Sep 21, 2026

Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
  • [INDIRECT_PROMPT_INJECTION]: The skill processes session logs which are considered untrusted data sources as they may contain external inputs or prior instructions.
  • Ingestion points: Reads local session logs and project files such as package.json or CI workflows as described in SKILL.md.
  • Boundary markers: No specific delimiters or instructions to disregard embedded commands in the ingested logs are provided.
  • Capability inventory: Capability is primarily limited to file reading and invoking a local writing style skill; it lacks network access or direct code execution functions that would significantly escalate the risk.
  • Sanitization: Content from logs is used for reasoning and reporting without explicit sanitization steps.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 21, 2026, 12:01 PM
Security Audit — agent-trust-hub — retro