fix-sentry

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

• Third-party content exposure detected (high risk: 0.90). The skill explicitly fetches and ingests Sentry data (see Phase 1 Step 1.2 mcp__sentry__list_issues and Step 1.5 mcp__sentry__get_issue_details) — untrusted, user-generated error messages and stack traces are parsed and used to triage and drive fixes/PR creation, so third-party content can materially influence agent actions.