to-issues

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

• Third-party content exposure detected (high risk: 0.75). The required workflow explicitly says that if the user passes an issue reference (URL/number/path), the agent will fetch the issue tracker and read the full body and comments, which are outsider-authored free text (issue descriptions/comments by others) and will be ingested into the LLM context via the issue-tracker fetch path.