vercel-incident-runbook

Warn

Audited by Gen Agent Trust Hub on Mar 12, 2026

Risk Level: MEDIUMCOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
  • [DATA_EXPOSURE]: The skill contains instructions to retrieve and decode sensitive Kubernetes secrets (kubectl get secret vercel-secrets -o jsonpath='{.data.api-key}' | base64 -d), exposing authentication keys within the triage process.\n- [PRIVILEGE_ESCALATION]: The runbook utilizes high-privilege kubectl operations to modify the production environment, including updating secrets and restarting services, presenting a risk of unauthorized infrastructure changes.\n- [INDIRECT_PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection by ingesting unsanitized data from external health check URLs and API responses.\n
  • Ingestion points: references/examples.md, references/immediate-actions-by-error-type.md\n
  • Boundary markers: None present\n
  • Capability inventory: kubectl (read/write access), curl (outbound network requests)\n
  • Sanitization: No sanitization or validation of data retrieved from external sources is performed.
Audit Metadata
Risk Level
MEDIUM
Analyzed
Mar 12, 2026, 12:57 AM