planning

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

• Third-party content exposure detected (medium risk: 0.65). Outsider free text can enter the LLM context via the “Research & Analysis” phase’s runtime GitHub/remote repository analysis (e.g., gh reading issues/PRs/discussions and repomix --remote <github-repo-url> fetching external repo content), which is outsider-authored prose that the agent will ingest into its context.