ip-clause-review

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

• Third-party content exposure detected (high risk: 0.85). The skill ingests the agreement text provided at runtime (file/Drive link/pasted text) and then reads it into the LLM context; if that agreement is authored by an outsider (e.g., vendor/counterparty), its clause text is outsider-authored free-form prose that the LLM analyzes.