pr-human-guide

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

• Third-party content exposure detected (high risk: 0.80). The skill explicitly fetches and ingests user-generated GitHub PR content (pr_title, pr_body, and full diff via gh pr view / gh pr diff) and feeds that untrusted PR body/diff into its analysis and decision-making (Step 2–3 and the <untrusted_pr_content> block) which can materially influence what the agent writes back to the PR.