land-and-deploy

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

• Third-party content exposure detected (medium risk: 0.30). In SKILL.md, the runtime workflow uses GitHub CLI to read PR content and check results (e.g., gh pr view and gh pr checks in Steps 1–3), so outsider-authored text from the PR (submitted by external users) is ingested without the agent first selecting a specific item beyond using the PR for the current branch.