revise-spec-from-review

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

• Third-party content exposure detected (medium risk: 0.30). The skill’s required runtime revision workflow ingests outsider-authored review text (e.g., evals/closure/*/input/review.txt) via the prompt that applies the complete review, and the agent must read/act on that free text before editing specs.