eval-readme

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

• Third-party content exposure detected (high risk: 0.90). SKILL.md:7-13 describes fetching a GitHub repo’s README at runtime (via GitHub API / raw.githubusercontent.com) and then passing that README text into the LLM for evaluation, which is outsider-authored free text.

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

• Potentially malicious external URL detected (high risk: 0.90). The skill explicitly fetches README content at runtime (e.g., "gh api repos/{owner}/{repo}/readme --jq .content | base64 -d" and raw.githubusercontent.com), which injects remote content into the agent's context and therefore directly controls prompts.