perf

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

• Third-party content exposure detected (high risk: 0.70). The skill's /perf audit action (in SKILL.md) runs Lighthouse via /browser-test against arbitrary URLs, meaning the agent fetches and interprets content from public websites which can influence audit outcomes and subsequent decisions (e.g., regressions, ship checklist), creating a route for indirect prompt injection.