detection-engineering

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

• Third-party content exposure detected (medium risk: 0.65). The skill’s runtime workflow can ingest outsider-authored free text from user-provided inputs such as threat intelligence reports, log samples, or issue/PR-like drafts (“process user-supplied content that may include log samples, detection rule drafts, threat intelligence reports”), which the agent then uses as LLM context to author Sigma/ADS.