update-with-comments

Pass

Audited by Gen Agent Trust Hub on Jun 26, 2026

Risk Level: SAFEPROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
  • [PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection because it reads and acts upon external, untrusted data from GitHub Pull Request comments.
  • Ingestion points: Fetches PR conversation and inline comments using 'gh pr view' and 'gh api' as described in SKILL.md.
  • Boundary markers: No specific delimiters or instructions are provided to the agent to treat comment content as untrusted or to ignore potential instructions embedded within them.
  • Capability inventory: The agent has the capability to modify code in the workspace, commit changes, and push to remote branches based on its interpretation of these comments.
  • Sanitization: There is no evidence of validation or filtering of the comment content before it influences code generation.
  • [COMMAND_EXECUTION]: The skill uses the GitHub CLI (gh) to perform repository operations using user-supplied parameters.
  • Evidence: Outline uses 'gh pr view ' and 'gh api' for data retrieval, and instructions include git operations like stash, commit, and push.
Audit Metadata
Risk Level
SAFE
Analyzed
Jun 26, 2026, 10:28 AM
Security Audit — agent-trust-hub — update-with-comments