review-spec

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

• Third-party content exposure detected (high risk: 0.70). The required workflow only reads the thread-root spec.md from docs/threads/<...>/ (outsider-authored content category: “body text of issues/PRs/wiki/discussion threads authored by outsiders”), and that readable prose is then used to generate findings in the agent’s LLM context.