k8s-opentelemetry-collector-troubleshooting
Pass
Audited by Gen Agent Trust Hub on Nov 8, 2026
Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill uses kubectl to modify the cluster state, including applying configuration manifests and restarting deployments in the remediation sections of SKILL.md.
- [INDIRECT_PROMPT_INJECTION]: The skill ingests untrusted data from the Kubernetes environment, creating a potential surface for indirect prompt injection.
- Ingestion points: The skill retrieves container logs and cluster events via kubectl logs and kubectl get events in SKILL.md and references/export-and-operator.md.
- Boundary markers: The instructions utilize specific grep and jsonpath filters to limit the data returned to the agent, providing a basic boundary against injection.
- Capability inventory: The skill is capable of executing kubectl apply and kubectl rollout restart across its main instructions and reference files.
- Sanitization: No explicit sanitization or validation of the ingested log or event content is performed before processing.
Audit Metadata