prompt-test

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

• Third-party content exposure detected (high risk: 0.75). The required workflow fetches outsider-authored PR content at runtime via gh api "repos/${REPO}/pulls/${N}" (public GitHub PR body/diff metadata), which the reviewer script then uses to generate LLM review text—creating an indirect prompt-injection path from outsider PR authors into the agent’s LLM context.