workflow-release-prep

Pass

Audited by Gen Agent Trust Hub on Oct 7, 2026

Risk Level: SAFEINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
  • [COMMAND_EXECUTION]: The skill executes multiple Git commands to manage the local environment and remote repository.
  • Evidence: Phase 0 and Phase 4 in SKILL.md use git status, git diff, git log, git branch, and git push to interact with the file system and remote origin.
  • [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted data from the local working tree which could contain malicious instructions intended to influence the agent's behavior.
  • Ingestion points: The skill reads local file changes, untracked files, and commit logs via git diff, git status --porcelain, and git log as described in Phase 0 of SKILL.md.
  • Boundary markers: There are no explicit delimiters or specific instructions to the agent to ignore instructions embedded within the code being reviewed.
  • Capability inventory: The skill can stage files (git add), push code to remote branches (git push), and trigger other tools like workflow-pr to open pull requests.
  • Sanitization: The skill incorporates defensive steps in Phase 1 and Phase 2, requiring the use of audit-code-review and plan-secrets-audit to inspect the aggregate diff for accidental inclusions or security risks before proceeding.
Audit Metadata
Risk Level
SAFE
Analyzed
Oct 7, 2026, 11:27 PM