code-porter

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

• Third-party content exposure detected (high risk: 0.90). The skill's required workflow (e.g., "Phase 2: 方案搜索" and the Quick Reference/Search Strategy) explicitly directs searching and evaluating open web sources like GitHub, npm/PyPI, Stack Overflow, Reddit and Twitter — untrusted, user-generated public content that the agent is expected to read and use to make library/implementation decisions.