qa-changes

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

• Third-party content exposure detected (high risk: 0.90). The SKILL.md explicitly instructs the agent to read PR diffs, titles, descriptions and any linked issues (Phase 1) and to base its hypothesis and subsequent test actions on that user-generated GitHub/issue content, which are untrusted third-party inputs that can materially influence tool use and next actions.