web-search-techniques

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

• Third-party content exposure detected (high risk: 1.00). The SKILL.md explicitly instructs the agent to perform web research with WebSearch/WebFetch and to fetch and process content from arbitrary public web pages (see "Scope: web-only", "WebFetch Guidance", and the "webfetch.py Fallback" that retrieves open URLs and the Source Authority Ranking that includes GitHub/StackOverflow), so untrusted third‑party content will be ingested and can influence tool use and decisions.