autoresearch

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

• Third-party content exposure detected (high risk: 1.00). The skill's SKILL.md research loop explicitly performs WebSearch and "WebFetch the page" for top results and the agent must extract and synthesize claims from those fetched public web pages (references/program.md further governs source preferences), so it ingests open, potentially user-generated third‑party content that can influence subsequent actions like filing wiki pages.