container-security
Pass
Audited by Gen Agent Trust Hub on May 13, 2026
Risk Level: SAFEEXTERNAL_DOWNLOADSCOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
- [EXTERNAL_DOWNLOADS]: Fetches container management tools (k9s, ArgoCD, yq, Kustomize) from official GitHub release pages. These references target a well-known service and include version pinning to ensure reproducibility and integrity.
- [COMMAND_EXECUTION]: Uses Azure CLI (az acr) and Trivy for container registry management and vulnerability scanning. These commands are necessary for the skill's intended purpose and operate within expected parameters.
- [PROMPT_INJECTION]: The skill ingests and processes external JSON vulnerability reports from Trivy.
- Ingestion points: Reads and parses
trivy-results.json(SKILL.md). - Boundary markers: No explicit delimiters or instructions are provided to the agent to ignore embedded commands within the scan data.
- Capability inventory: The skill has access to powerful tools including
Bash,Write,Edit, andAgent(SKILL.md). - Sanitization: The provided Python parsing logic lacks sanitization or validation of the input fields beyond simple field extraction.
Audit Metadata