release-openclaw-maintainer

Pass

Audited by Gen Agent Trust Hub on Jul 27, 2026

Risk Level: SAFEPROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
  • [PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection (Category 8) because it ingests external, untrusted data to generate changelogs and release notes.
  • Ingestion points: The skill reads git commit history (git log) and pull request metadata (gh api) from the OpenClaw repository.
  • Boundary markers: There are no explicit instructions for the model to ignore embedded commands within the ingested data, although the skill enforces a strict output format.
  • Capability inventory: The skill possesses extensive capabilities, including shell command execution (pnpm, git, gh), file modifications, and GitHub Actions dispatching.
  • Sanitization: The instructions require a human maintainer to review and edit the generated changelog before it is committed, providing a layer of human-in-the-loop sanitization.
  • [COMMAND_EXECUTION]: The skill executes a wide range of shell commands to manage the release lifecycle. While these are necessary for its intended purpose, they provide a large attack surface if the agent is manipulated via prompt injection.
  • [CREDENTIALS_UNSAFE]: The skill manages sensitive credentials for NPM and API keys through private documentation and context variables ($release-private). It follows security best practices by instructing the agent to use tmux and avoid printing secrets to the terminal transcript to prevent accidental exposure.
Audit Metadata
Risk Level
SAFE
Analyzed
Jul 27, 2026, 02:30 PM
Security Audit — agent-trust-hub — release-openclaw-maintainer