openleash

Name: openleash
Availability: InStock
Author: OpenLeash

Authorize AI agent actions locally with verifiable PASETO tokens.

6|Updated Feb 16, 2026

Search Tags:#authorization #local-first #agent-management #ai-security #ed25519 #policy-gate #paseto

Ready to run

npx skills add https://github.com/openleash/openleash.ai --skill openleash

💬 Prompt for ChatGPT / Claude Chat▼

Please help me install this Agent Skill.
Skill: openleash
Source: https://github.com/openleash/openleash.ai/tree/main/public
Command: npx skills add https://github.com/openleash/openleash.ai --skill openleash

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

OpenLeash provides local-first authorization for AI agents, ensuring risky actions are gated by verifiable proofs before execution.

Core Features & Use Cases

Local authorization alongside AI runtimes to approve actions before they incur side-effects.
Issues verifiable PASETO v4.public tokens and supports Ed25519 signing for requests.
Fully documented SDKs (TypeScript, Python, Go) and a built-in web GUI for management.

Quick Start

Install the OpenLeash SDK, initialize your agent, and start the local policy sidecar to begin authorizing actions.

Frequently Asked Questions about openleash

High-intent search queries and answers about installing and using this skill.

FAQPage Schema

How do I authorize AI agent actions locally before they execute risky operations?▼

To authorize AI agent actions locally, you use a local-first policy gate that evaluates requests and issues verifiable proofs before side-effects occur. OpenLeash provides this by running a local server to gate risky operations.

How does Ed25519 signing secure local AI authorization workflows?▼

Ed25519 signing secures local AI authorization by cryptographically signing action requests, which are then verified before execution. OpenLeash uses Ed25519 keys to ensure requests originate from authenticated agents.

Do I need Node.js running to use local-first policy evaluation for AI security?▼

Yes, local-first policy evaluation requires Node.js to run. You must have the OpenLeash server executing locally at http://127.0.0.1:8787 to evaluate policies and issue PASETO tokens.

What is a PASETO v4.public token used for in AI agent management?▼

A PASETO v4.public token in AI agent management provides an offline-verifiable proof of authorization. OpenLeash issues these tokens to prove an agent action was approved before execution.

Can I integrate local authorization using TypeScript, Python, and Go SDKs?▼

Yes, you can integrate local authorization using fully documented TypeScript, Python, and Go SDKs. These SDKs interface with the local OpenLeash server to manage agent policies and tokens.

When should I use a local policy sidecar over other AI security approaches?▼

You should use a local policy sidecar when you need on-demand authorization with offline proof verification for actions like purchases or messages. OpenLeash gates these side-effects locally without relying on external cloud services.