goosetown-researcher-stackoverflow

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

• Third-party content exposure detected (high risk: 0.90). This skill explicitly fetches and ingests user-generated Q&A from Stack Overflow via the Stack Exchange API (https://api.stackexchange.com/2.3/ — e.g., using ?filter=withbody and /questions/.../answers) and requires the agent to read and synthesize those answer bodies into research briefs, so untrusted third‑party content can directly influence the agent's outputs and behavior.