troubleshoot-ssi
Pass
Audited by Gen Agent Trust Hub on Sep 15, 2026
Risk Level: SAFEEXTERNAL_DOWNLOADSCOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [EXTERNAL_DOWNLOADS]: The skill downloads the
puptool from the officialdatadog-labsGitHub repository for Linux environments. This is an expected operational dependency for a skill authored by the same vendor.\n- [INDIRECT_PROMPT_INJECTION]: The skill consumes external data like logs and pod manifests, creating a potential surface for injection if an attacker controls those inputs.\n - Ingestion points: Reads
kubectl logs,kubectl describe podoutput, andpuptroubleshooting telemetry.\n - Boundary markers: None identified for delimiting ingested data.\n
- Capability inventory: Includes
kubectl rollout restart,kubectl apply,kubectl label, andkubectl exec.\n - Sanitization: No specific validation or sanitization of the ingested logs or metadata is described before processing.\n- [COMMAND_EXECUTION]: The skill performs administrative actions on the Kubernetes cluster via
kubectl, which is necessary for the intended troubleshooting and remediation flow.
Audit Metadata