sglang-prod-incident-triage

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

• Third-party content exposure detected (high risk: 0.80). The skill's required workflow and scripts (e.g., scripts/incident_artifact_tool.py collect-bundle which fetches arbitrary base_url endpoints like /health, /metrics, /v1/loads and the dump-handling paths that pickle-load request/crash .pkl files used by summarize-dump and scripts/replay_trusted_request_dump.py) ingest untrusted, potentially user-generated content and then parse it to produce signals and next-step triage decisions, so third-party content can materially influence agent actions.