fusion-github-review-resolution

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

• Third-party content exposure detected (high risk: 1.00). The skill explicitly fetches and ingests user-generated GitHub pull request review threads (see SKILL.md steps 2–3 and assets/pull-request-review-threads.graphql plus scripts/get-review-comments.sh and resolve-review-comments.sh which call gh api graphql), requires the agent to read and interpret those comments as part of the fix/reply/resolve workflow, and uses them to decide and perform code changes — so untrusted third-party review content can materially influence agent actions.