research-team

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

• Third-party content exposure detected (high risk: 0.85). Outsider free text can enter the LLM context via runtime web content fetched during the worker’s research phases (e.g., protocols/market-analysis.md, competitive-analysis.md, academic-research.md, stack-evaluation.md, research.md all instruct web searching and then synthesize results), which typically includes arbitrary page text authored by third parties.