to-issues

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

• Third-party content exposure detected (high risk: 0.85). The skill’s runtime path explicitly fetches a user-supplied PRD GitHub issue via gh issue view <number> --comments, which ingests outsider-authored free text from the PRD issue body/comments into the agent’s LLM context (unless the PRD was authored by the operating user, which is not guaranteed).