skills/zernie/vigiles/review-docs/Gen Agent Trust Hub

review-docs

Pass

Audited by Gen Agent Trust Hub on Jul 20, 2026

Risk Level: SAFEPROMPT_INJECTION
Full Analysis
  • [PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection because it processes untrusted documentation content without proper isolation.
  • Ingestion points: The skill reads documentation files (defaulting to README.md) from a path provided by the user.
  • Boundary markers: The instructions do not use specific delimiters (such as XML tags or triple backticks) to separate the content of the document being reviewed from the agent's instructions.
  • Capability inventory: The skill primarily generates analytical reports but can be instructed to modify files if the optional "iterate to a target" loop is activated.
  • Sanitization: No sanitization or filtering of the document content is performed to prevent embedded instructions from influencing the subagents' evaluations or the synthesis phase.
Audit Metadata
Risk Level
SAFE
Analyzed
Jul 20, 2026, 04:44 PM
Security Audit — agent-trust-hub — review-docs