fix-security-vulnerability

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

• Third-party content exposure detected (high risk: 1.00). The SKILL.md explicitly fetches Dependabot alerts from GitHub via gh api repos/getsentry/sentry-javascript/dependabot/alerts... (see "Scan Step 1: Fetch All Open Alerts" and "Step 1: Fetch Vulnerability Details") and requires the agent to read and act on the API response (alerts/descriptions) to decide fixes, branches, or dismissals, which exposes it to untrusted third-party alert content that could contain injected instructions.