research-spike

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

• Third-party content exposure detected (high risk: 0.85). The required workflow explicitly fetches “current documentation via Context7” (e.g., npx ctx7@latest ... and npx ctx7@latest docs ...), which at runtime ingests outsider-authored public/vendor documentation text into the agent’s LLM context.

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

• Potentially malicious external URL detected (high risk: 1.00). The skill explicitly runs "npx ctx7@latest ..." during runtime, which fetches and executes remote code from the npm registry (e.g. https://www.npmjs.com/package/ctx7), so it relies on an external package that executes code at runtime.