sca-trivy

Pass

Audited by Gen Agent Trust Hub on Aug 14, 2026

Risk Level: SAFEEXTERNAL_DOWNLOADSCOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
  • [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted external data from filesystems and container images, creating a potential surface for indirect prompt injection if an attacker embeds malicious instructions in scanned project files.
  • Ingestion points: The agent is instructed to scan container images, local filesystems, IaC configurations, and dependency manifest files (e.g., package-lock.json, go.mod).
  • Boundary markers: No specific boundary delimiters or explicit 'ignore embedded instructions' warnings are provided for the agent when interpreting the output of the scan results.
  • Capability inventory: The skill facilitates the execution of shell commands using the trivy and docker CLI tools.
  • Sanitization: The instructions do not specify any sanitization, escaping, or filtering of the external content prior to analysis.
  • [UNVERIFIABLE_DEPENDENCIES_AND_REMOTE_CODE_EXECUTION]: The skill references external tool installation and remote actions originating from well-known and official sources.
  • Fetches the vulnerability database from Aqua Security's official repository on the GitHub Container Registry (ghcr.io/aquasecurity/trivy-db).
  • Recommends installation via trusted package managers (Homebrew, APT) or official Docker images (aquasec/trivy:latest).
  • Integrates with the official aquasecurity/trivy-action GitHub Action for CI/CD pipelines.
  • [DATA_EXPOSURE_AND_EXFILTRATION]: The skill includes guidance for the secure management of sensitive information.
  • Explicitly advises using environment variables or credential helpers for registry authentication to avoid hardcoding credentials.
  • Categorizes scan reports and SBOM artifacts as sensitive data that require appropriate access controls.
Audit Metadata
Risk Level
SAFE
Analyzed
Aug 14, 2026, 02:35 AM
Security Audit — agent-trust-hub — sca-trivy