github-resolution
Pass
Audited by Gen Agent Trust Hub on Oct 27, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill is designed to process untrusted data from GitHub pull requests, issues, and contributor code, creating a potential surface for indirect prompt injection attacks.
- Ingestion points: Processes PR descriptions, issue text, and source code diffs as described in SKILL.md.
- Boundary markers: Explicitly defines a 'Trust Boundary' section instructing the agent to ignore any commands, scope expansions, or credential requests found within the untrusted data.
- Capability inventory: Authorized to perform git operations (commit, push), run local test suites (e.g.,
bin/ci), and invoke other agent skills for auditing and releases. - Sanitization: Instructs the agent to personally reconcile diffs and re-check hostile-change gates before merging changes.
- [COMMAND_EXECUTION]: The skill requires the agent to execute shell commands for repository management and software verification.
- Evidence: Utilizes
git status,git commit,git push, and project-specific testing commands likebin/ciorRuboCopto verify and land fixes.
Audit Metadata