incident-runbook-templates
Pass
Audited by Gen Agent Trust Hub on Sep 26, 2026
Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill documentation includes code blocks containing shell commands for Kubernetes management (
kubectl), database administration (psql), and network requests (curl). These are presented as reference templates for incident response procedures. - [INDIRECT_PROMPT_INJECTION]: The templates outline processes for responding to external data sources, including alerts from PagerDuty/Slack, application logs, and monitoring metrics. This establishes an attack surface where untrusted external data could attempt to influence an agent's response actions during an incident.
- Ingestion points: Alert notifications (PagerDuty/Slack), application logs (
kubectl logs), and monitoring metrics (Prometheus/Grafana). - Boundary markers: None identified in the templates to separate untrusted data from the agent's instructions.
- Capability inventory: Container orchestration (
kubectl), database query execution (psql), and network communication (curl). - Sanitization: The templates do not specify validation or sanitization requirements for external data processed during triage.
Audit Metadata