Documentation
Everything you need to start enforcing capability policies on your AI agents.
Getting Started
Install wsh and enforce your first policy in under two minutes.
Integrating with AI Agents
Run agents through wsh transparently with warrant-box, shell replacement, wrappers, and more.
Concepts
Manifests, drafts, lock compilation, enforcement, and the architecture.
CLI Reference
Every wsh command, flag, and option documented.
Manifest Specification
Open StandardThe open standard for declaring CLI tool capabilities. Schema definition, CLI control surface taxonomy, registry protocol, bundle format, and validation rules. CC0 licensed — implement freely.
Manifest Authoring Guide
Practical guidance for writing manifests, with examples from git, cargo, npm, and more.
Security Model
Threat model, honest trade-offs, and what wsh does and doesn't protect against.