fix-bug

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

• Third-party content exposure detected (high risk: 0.80). The workflow explicitly instructs the agent to perform external web research and web navigation (Phase 1 Agent 3: "Search ... using Context7 and web search"), to fetch issue data (examples show gh issue view), and to use the agent-browser skill to navigate public pages—all of which ingest untrusted, user-generated or public web content that can influence debugging decisions and actions.