skills/cerbos/skills/cerbos-hub-setup/Gen Agent Trust Hub

cerbos-hub-setup

Pass

Audited by Gen Agent Trust Hub on Nov 1, 2026

Risk Level: SAFEEXTERNAL_DOWNLOADSCOMMAND_EXECUTION
Full Analysis
  • [EXTERNAL_DOWNLOADS]: The skill's scripts utilize npx -y cerbosctl to dynamically download and run the official Cerbos CLI if it is not present on the system path. This targets a well-known registry (NPM) for the vendor's own tooling.
  • [EXTERNAL_DOWNLOADS]: Deployment instructions pull official Cerbos Docker images from the GitHub Container Registry (ghcr.io/cerbos/cerbos).
  • [COMMAND_EXECUTION]: The skill executes local shell scripts and vendor-specific CLI tools (cerbosctl, docker) to manage policy stores and PDP instances.
  • [CREDENTIALS_UNSAFE]: While the skill manages sensitive API secrets (CERBOS_HUB_CLIENT_SECRET), it implements best-practice safeguards. It explicitly instructs the agent to never output secrets, provides commands for the user to securely set environment variables (using read -rs), and includes logic in scripts/check-prereqs to verify if a secret is set while explicitly withholding its value from the output.
  • [SAFE]: No malicious patterns, prompt injections, or unauthorized data exfiltration attempts were detected. The skill is consistent with its stated purpose of managing Cerbos Hub infrastructure.
Audit Metadata
Risk Level
SAFE
Analyzed
Nov 1, 2026, 06:46 PM