review

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

• Third-party content exposure detected (medium risk: 0.65). The workflow ingests outsider-authored free text from the repo at runtime—e.g., issue/PR/spec text referenced via commit messages and PRD/spec files under docs/, specs/, .scratch/, plus standards files like CLAUDE.md/AGENTS.md/CONTRIBUTING.md/ADRs that may be authored by others—then passes that readable text into the sub-agents’ LLM context.