search-strategy

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

• Third-party content exposure detected (high risk: 0.85). This skill’s required workflow performs source-specific searches across connected systems (e.g., ~~chat, ~~knowledge base/wiki, ~~project tracker) and then “synthesize[s] the results,” which implies the LLM ingests outsider-authored free text from those sources at runtime (e.g., other users’ chat messages, wiki pages, or tracker comments/threads).