qa

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

• Third-party content exposure detected (high risk: 0.80). The SKILL.md explicitly instructs using scripts/skills-browser.sh to "navigate " and "read" browser state (Browser Capture section), meaning the agent will fetch and interpret content from arbitrary URLs as part of its QA workflow, which could include untrusted/public pages and thus enable indirect prompt injection.