policy-and-managed-settings

Pass

Audited by Gen Agent Trust Hub on Aug 10, 2026

Risk Level: SAFE
Full Analysis
  • Policy Management Architecture: The skill provides detailed instructions on how enterprise policies are declared, exported, and enforced. This includes a clear flowchart for determining where a policy should reside (SDK/runtime vs. VS Code core), which is a standard security best practice for maintaining authoritative enforcement points.
  • External Configuration Ingestion: The documentation describes how 'managed settings' are delivered via native MDM (Windows registry/macOS plist), GitHub server endpoints, and local files. The skill outlines a 'most-restrictive-wins' precedence model and emphasizes the use of a shared normalizer (normalizeManagedSettings) and projection logic (projectManagedSettings) to validate and type-check external input before it is applied to the agent's configuration.
  • Development and Testing Tools: The skill references local development tools, including an npm run mock-policy-server for simulating policy responses and an export-policy-data script for generating catalog artifacts. These are standard engineering practices for verifying policy behavior in isolated environments.
  • Security Boundary Documentation: The guides explicitly discuss the separation of concerns between sandbox containment and permission authorization, ensuring that policy state and execution environments remain aligned without expanding access unexpectedly.
  • Vendor-Specific Resources: The skill refers to internal Microsoft and VS Code repositories and packages, such as @vscode/policy-watcher. These references are consistent with the skill's purpose as a guide for developers working on the official product suite.
Audit Metadata
Risk Level
SAFE
Analyzed
Aug 10, 2026, 08:39 PM
Security Audit — agent-trust-hub — policy-and-managed-settings