outside-to-issue

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

• Third-party content exposure detected (high risk: 0.75). The required workflow reads .md files from ai-docs/reviews-pr-<PR>/outside/ and injects their free-form review text into newly generated issues/*.md content that is then used by the agent/LLM, and those outside/ review files are explicitly “outside-of-diff” (i.e., not authored by the operating user/agent).