code-comments

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

• Third-party content exposure detected (high risk: 0.70). The required workflow can ingest outsider-authored free text from the contents of files the user selects for review (e.g., PR/issue/comment bodies or diff text) via the extractor’s runtime reading of those files’ comment text and surrounding code context, which the agent then includes in LLM context.