to-issues

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

• Third-party content exposure detected (high risk: 0.75). At runtime, step 1 says that if the user provides an issue reference (URL/path/number), the skill fetches and reads the full issue正文和评论 from the issue tracker; those issue/comment texts are outsider-authored free-form content that can be ingested into the agent’s LLM context.