trivy
Pass
Audited by Gen Agent Trust Hub on Sep 3, 2026
Risk Level: SAFECOMMAND_EXECUTIONEXTERNAL_DOWNLOADSINDIRECT_PROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill provides instructions and syntax for executing the
trivycommand-line utility. It supports scanning container images (trivy image), local directories (trivy fs), and Kubernetes cluster configurations (trivy k8s). - [EXTERNAL_DOWNLOADS]: The skill facilitates the retrieval of container images from public registries (e.g., Docker Hub) for analysis. These downloads are directed to well-known services as part of the tool's standard functionality.
- [INDIRECT_PROMPT_INJECTION]: The skill processes external, untrusted artifacts such as container images and filesystem contents. This creates a surface where findings from these artifacts could potentially include malicious instructions intended to influence the agent's behavior during triage.
- Ingestion points: Input targets such as container image tags, local filesystem paths, and Kubernetes manifests.
- Boundary markers: Absent; the agent is instructed to triage findings directly from the tool's output.
- Capability inventory: The skill utilizes subprocess execution to run the
trivybinary. - Sanitization: No specific sanitization or filtering of the external data is defined within the skill's instructions.
Audit Metadata