spec

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

• Third-party content exposure detected (high risk: 0.70). The required workflow loads thread context by reading seed.md, decisions.md, optional proposal.md, and a referenced artifact/prompt (which can include outsider-authored free text such as issue/PR/wiki content), and that readable text is then used to draft spec.md—so the LLM context can be fed outsider text via the “Load context” step.