alert-investigation

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

• Third-party content exposure detected (medium risk: 0.30). In SKILL.md, the runtime workflow reads alert details, performs knowledge-graph searches, and queries tracing/metrics/log patterns (e.g., “Search for the specific alert…”, “Search the knowledge graph…”, and “Query for specific log patterns or error messages”) using data that can include outsider-authored free text such as user-generated log messages and alert/monitor labels.