pod-deploy
Pass
Audited by Gen Agent Trust Hub on Oct 13, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill instructs the agent to read container boot logs via the
stream-pod-logstool to confirm an application has successfully bound to a port. - Ingestion points: Container boot logs (SKILL.md).
- Boundary markers: None specified, though instructions focus on matching specific "listening" strings.
- Capability inventory:
create-pod,delete-pod,pod-action,create-network-volume,update-ssh-keys. - Sanitization: None specified for the log stream.
- [EXTERNAL_DOWNLOADS]: The skill handles the selection and deployment of container images. It includes logic to resolve images from templates, Hub repositories, or existing pods, ensuring version pinning and CUDA compatibility before execution. These operations are performed through official provider infrastructure.
Audit Metadata