diagnose-and-fix-tracked-errors

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

• Third-party content exposure detected (high risk: 0.70). This skill is a methodological prompt that directs the agent to “pull the full context from the observability tool” (including stack traces/logs and captured variables) and paste them into the LLM context, where that tool data originates from runtime systems and may include outsider-authored free text (e.g., request contents, user messages embedded in logs).