address-review

Pass

Audited by Gen Agent Trust Hub on Oct 1, 2026

Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
  • [INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and act upon code review findings from external sources such as GitHub Pull Requests and Issues.
  • Ingestion points: The agent retrieves review bodies using gh pr view or gh issue view and stores them in /tmp/ files in SKILL.md before processing them with the provided Python script.
  • Boundary markers: There are no explicit boundary markers or instructions to the agent to treat external content as untrusted. The skill explicitly instructs the agent to treat the 'Done when' field as a 'contract' and to 'Run the check the Done when names'.
  • Capability inventory: The agent has the capability to modify project source code, execute build/check commands within the project environment, commit and push changes via Git, and interact with the GitHub API to post comments and edit PR/Issue metadata.
  • Sanitization: The parsing logic in scripts/open_findings.py extracts content from the review round but does not validate or sanitize the instructions contained within the 'Fix', 'Done when', or 'Question' fields before the agent processes them.
Audit Metadata
Risk Level
SAFE
Analyzed
Oct 1, 2026, 07:58 AM