pr-review-workflow

MEDIUM W011: Third-party content exposure detected (indirect prompt injection risk).

• Third-party content exposure detected (high risk: 0.90). SKILL.md explicitly requires the agent to read and act on GitHub PR review threads and notifications (e.g., selecting which reviewers to re-request, resolving threads, using GitHub API endpoints), which are user-generated/untrusted third-party contents that the agent must interpret and that can materially change its actions.

MEDIUM W012: Unverifiable external dependency detected (runtime URL that controls agent).

• Potentially malicious external URL detected (high risk: 1.00). The skill's session gate requires running compose-agentsmd which (per agent-ruleset.json "source": "github:metyatech/agent-rules@HEAD") will fetch rules from the external GitHub repo at github:metyatech/agent-rules@HEAD at runtime, and those fetched rules directly control the agent's AGENTS.md instructions/prompts, making this a required runtime external dependency.