gh-credit-harvest

Pass

Audited by Gen Agent Trust Hub on Jun 25, 2026

Risk Level: SAFECOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
  • [COMMAND_EXECUTION]: The skill executes development tools including cargo test and a local utility script scripts/check-coauthor-trailers.py found in the repository. These commands are used to verify the integrity and formatting of code from Pull Requests before they are committed.
  • [PROMPT_INJECTION]: The skill has an attack surface for indirect prompt injection as it processes content from external GitHub Pull Requests.
  • Ingestion points: Pull Request code diffs, comments, and issue descriptions are read into the agent's context using gh pr view and gh pr diff commands.
  • Boundary markers: The instructions include explicit directives to "Review from evidence, not the title" and to "Read the diff, tests, linked issue, comments, and CI," which serves as a prompt-level boundary for the agent.
  • Capability inventory: The agent has the ability to execute shell commands (git, gh, cargo, python3), modify the local repository state, and perform write operations on GitHub (posting comments).
  • Sanitization: The skill relies on the agent's internal evaluation and manual review of the fetched content rather than automated sanitization filters.
Audit Metadata
Risk Level
SAFE
Analyzed
Jun 25, 2026, 12:01 PM
Security Audit — agent-trust-hub — gh-credit-harvest