receiving-code-review

Pass

Audited by Gen Agent Trust Hub on Sep 21, 2026

Risk Level: SAFEINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
  • [INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and act upon feedback from external code reviews, which is an untrusted data source that could be used for indirect prompt injection.
  • Ingestion points: Review comments from external contributors or partners (SKILL.md).
  • Boundary markers: The skill lacks explicit markers to distinguish review content as data, relying on the agent's technical evaluation to detect issues.
  • Capability inventory: The skill uses grep for codebase analysis, gh api for GitHub interactions, and general project file modification capabilities.
  • Sanitization: No automated sanitization or escaping of feedback text is implemented before processing.
  • [COMMAND_EXECUTION]: The skill executes standard command-line tools for development tasks.
  • Evidence: Uses grep to search the codebase for feature usage and gh api to submit replies to specific GitHub pull request comment threads (SKILL.md).
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 21, 2026, 01:38 AM
Security Audit — agent-trust-hub — receiving-code-review