doc-skills-md

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

• Third-party content exposure detected (high risk: 0.90). The skill explicitly fetches and reads a marketplace catalog (e.g., the public GitHub repo MOlechowski/claude-skills via "gh api repos/.../marketplace.json" or a .claude-plugin/marketplace.json) and uses that untrusted, third‑party plugin data to decide which plugins to recommend/install, so external content can influence agent behavior.

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

• Potentially malicious external URL detected (high risk: 0.70). The skill instructs at runtime to fetch the marketplace catalog (e.g., gh api repos/MOlechowski/claude-skills/contents/.claude-plugin/marketplace.json and https://github.com/MOlechowski/claude-skills) and uses that remote JSON to drive which plugins are recommended, so external content directly influences the agent's output.