Universal hook engine for AI coding tools. Write policy once — enforce it everywhere, or let external systems drive it.

ai-hooks is an open-source framework for portable guardrails, audit trails, and runtime control across AI coding tools. Define hooks in one config and they work everywhere — Claude Code, Codex, Gemini CLI, Cursor, Kiro, Cline, and more. Or use it as a library to let external platforms programmatically steer the AI client a developer is using.

AI coding tools are proliferating, and every team hits the same wall:

• No unified policy layer. Each tool has its own hook format, so guardrails get duplicated, drift, and break silently.
• No external control plane. If you need to programmatically steer the client a developer is using — set boundaries, collect signals, enforce workflow rules — there's no standard way to do it. You're locked into one vendor's API.
• Fragmented audit trails. Shell commands, file writes, and tool calls are logged in different formats per tool, making compliance and debugging painful.
• Inconsistent enforcement. Some tools can block dangerous actions before they happen; others can only observe after the fact. You need a framework that normalizes this.

ai-hooks gives you a single policy engine that can steer any AI coding client with native hook support. Write rules once, and adapters translate them into the exact config format each tool understands. Block, observe, or redirect actions across every tool from one place.

graph TD
    Config["<b>ai-hooks.config.ts</b><br/>Define once"]
    Engine["Hook Engine<br/><i>middleware chain, priority, filters</i>"]
    Platform["External Platform<br/><i>orchestrator, PM tool, CI system</i>"]

    Config --> Engine
    Platform -->|"library API"| Engine

    Engine -->|"adapters generate<br/>native configs"| native

    subgraph native ["Hook-compatible clients — blocking + observation"]
        CC["Claude Code"]
        Codex["Codex CLI"]
        Gemini["Gemini CLI"]
        Cursor["Cursor"]
        Kiro["Kiro"]
        OC["OpenCode"]
        Droid["Factory Droid"]
        Amp["Amp"]
        Cline["Cline"]
    end

    style Config fill:#7c3aed,color:#fff,stroke:none
    style Engine fill:#2563eb,color:#fff,stroke:none
    style Platform fill:#059669,color:#fff,stroke:none
    style native fill:#f0fdf4,stroke:#16a34a

Two ways to use it:

1. Config-driven — Write an ai-hooks.config.ts, run ai-hooks generate, and adapters produce native hook configs for every detected tool. When the agent tries to run a shell command, write a file, or call a tool, your hooks fire natively inside the client — blocking dangerous actions, scanning for secrets, and logging everything before it happens.

2. Library / control plane — Import the engine into your platform. Programmatically steer the AI client a developer is using — set boundaries, collect signals, and enforce workflow rules — all through a standard API that works regardless of which tool they're running.

# Install
npm i -D @premierstudio/ai-hooks

# Initialize config
npx ai-hooks init

# Detect installed tools and generate hooks
npx ai-hooks detect --verbose
npx ai-hooks generate
npx ai-hooks install

// ai-hooks.config.ts
import { defineConfig, hook, builtinHooks } from "@premierstudio/ai-hooks";

export default defineConfig({
  extends: [{ hooks: builtinHooks }],
  hooks: [
    hook("before", ["shell:before"], async (ctx, next) => {
      if (ctx.event.command.includes("npm publish")) {
        ctx.results.push({ blocked: true, reason: "Publishing is restricted" });
        return;
      }
      await next();
    })
      .id("org:block-publish")
      .name("Block npm publish")
      .priority(10)
      .build(),
  ],
  settings: {
    hookTimeout: 5000,
    failMode: "open",
    logLevel: "warn",
  },
});

ai-hooks ships with 4 hooks you can use immediately:

15 event types across the full AI tool lifecycle:

If you're building a platform that orchestrates AI agents — a PM tool, CI system, or custom control plane — import the engine directly:

import { HookEngine, builtinHooks } from "@premierstudio/ai-hooks";

const engine = new HookEngine({
  hooks: builtinHooks,
  settings: { failMode: "closed" },
});

// Check if an action should be blocked
const result = await engine.isBlocked(event, { name: "my-platform", version: "1.0" });
if (result.blocked) {
  console.log(`Blocked: ${result.reason}`);
}

This gives your platform a standard interface for controlling what AI agents can do, regardless of which tool (Claude Code, Codex, Cursor, etc.) the developer is running.

git clone https://github.com/PremierStudio/ai-hooks.git
cd ai-hooks
npm install
npm run check   # lint + format + typecheck + test

See CONTRIBUTING.md for adding adapters and submitting PRs.

MIT