Skills
skills/jeremylongshore/claude-code-plugins-plus-skills/deepgram-incident-runbook/Gen Agent Trust Hub

deepgram-incident-runbook

Fail

Audited by Gen Agent Trust Hub on Feb 18, 2026

Risk Level: CRITICAL
Full Analysis
  • [COMMAND_EXECUTION] (LOW): The skill uses kubectl and curl for diagnostic tasks. These are powerful tools but are expected in a DevOps runbook context.
  • [DATA_EXPOSURE] (SAFE): No hardcoded credentials were found; the skill uses environment variables for authentication.
  • [INDIRECT_PROMPT_INJECTION] (LOW): Potential surface through external data ingestion. Ingestion points: Deepgram status API and local Prometheus metrics. Boundary markers: None. Capability inventory: curl, kubectl, and file-write. Sanitization: Limited to jq field extraction.
  • [EXTERNAL_DOWNLOADS] (SAFE): External links are to official domains. The 'status.in' malicious URL alert is a false positive.
Recommendations
  • Contains 1 malicious URL(s) - DO NOT USE
Audit Metadata
Risk Level
CRITICAL
Analyzed
Feb 18, 2026, 09:18 PM