commit-push-close
Pass
Audited by Gen Agent Trust Hub on Oct 17, 2026
Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTIONCREDENTIALS_UNSAFE
Full Analysis
- [COMMAND_EXECUTION]: The skill uses standard development tools including git, gh (GitHub CLI), pnpm, and composer to manage the shipping workflow. These commands are executed within a verified task checkout and are gated by combined user approval before any write actions occur.
- [INDIRECT_PROMPT_INJECTION]: The skill ingests untrusted data from issue bodies, comments, and diffs (Ingestion points: SKILL.md, references/ship-policy.md). It employs strong boundary markers by instructing the agent to treat repository content as evidence rather than instructions (Boundary markers: references/ship-policy.md). The toolset includes capabilities for repository mutation and external communication via GitHub (Capability inventory: SKILL.md). It enforces sanitization by scanning inputs for credentials and debug tags (Sanitization: references/ship-policy.md).
- [CREDENTIALS_SAFE]: The skill implements proactive safety measures by scanning staged changes and drafts for credential-shaped patterns (e.g., ghp_ tokens, private keys) and strictly forbids the staging of sensitive files like .env or .pem files.
Audit Metadata