plugin-search

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

• Third-party content exposure detected (high risk: 0.75). The skill downloads and parses outsider-authored marketplace JSON from public GitHub repos at runtime (via download_marketplace() calling the GitHub Contents API and decoding .claude-plugin/marketplace.json), then formats fields like description/homepage into plain-text output that can be ingested into the agent/LLM context.