groq-incident-runbook
Pass
Audited by Gen Agent Trust Hub on Sep 9, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill provides an automated evidence collection script that ingests external application data.\n
- Ingestion points: The
kubectl logscommand inreferences/communication-and-postmortem.mdingests logs from active deployments for analysis.\n - Boundary markers: The skill does not implement explicit boundary markers or delimiters when the agent reads or processes the ingested log data.\n
- Capability inventory: The skill is authorized to use
kubectlfor log retrieval andcurlfor external network operations to diagnostic endpoints.\n - Sanitization: The script includes a redaction step using
sedto filter out potential Groq API keys (gsk_tokens) before the logs are saved to the evidence bundle.
Audit Metadata