recipe-front-design

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

• Third-party content exposure detected (medium risk: 0.65). Non-user “external resource hearing” via external-resource-context (Step 4) can fetch/read public/other-party content into external-resources.md, which is then referenced by later agent prompts (e.g., Step 6 reads docs/project-context/external-resources.md), creating an indirect prompt-injection path.