merciagents

merciagents / riphook

Public

Deterministic security layer for Openclaw(Clawdbot), Cursor and Claude Code. Write secure code, prevent data exfil, and more

ai-agents clawdbot hooks moltbook moltbot
43
1
100% credibility
Found Feb 05, 2026 at 24 stars -- GitGems finds repos before they trend. Get early access to the next one.
Sign Up Free
AI Analysis
TypeScript
AI Summary

Riphook is a safety layer for AI coding tools that detects secrets and personal info, blocks dangerous commands, runs code checks, and logs activity for review.

How It Works

1
🔍 Discover Riphook

You hear about a helpful safety tool that keeps your AI coding buddy from making risky moves or spilling secrets.

2
📥 Run the quick setup

Copy and paste one simple command into your terminal, and it gently adds safety nets to your favorite coding helper.

3
🛡️ Safety shields activate

Your coding sessions now have invisible guards that watch for dangers and log every step, making everything feel secure.

4
💬 Chat with your AI as usual

Ask your AI to read files, run commands, or edit code just like before, but now with extra protection.

5
Spot a risky request
✅
Safe path

Good actions go through smoothly, keeping your work protected.

🚫
Blocked safely

Risky ones are halted with a friendly warning, so no harm done.

6
📊 Check your activity log

Peek at the simple log file to see a clear picture of what your AI did, like a helpful diary.

🎉 Code worry-free

Enjoy secure, traceable AI help that blocks mishaps and lets you focus on creating great code.

Sign up to see the full architecture

5 more

Sign Up Free

Star Growth

See how this repo grew from 24 to 43 stars Sign Up Free
Repurpose This Repo

Repurpose is a Pro feature

Generate ready-to-use prompts for X threads, LinkedIn posts, blog posts, YouTube scripts, and more -- with full repo context baked in.

Unlock Repurpose
AI-Generated Review

What is riphook?

Riphook is a TypeScript security layer for AI coding agents in Cursor, Claude Code, and OpenClaw/Clawdbot. It intercepts tool calls to block dangerous shell commands, scans prompts/files/outputs for secrets and PII to prevent data exfil, and runs static analysis on edited code via Semgrep. Install with a one-liner curl script that auto-configures hooks and emits Cursor agent-trace logs for visualizing agent flows.

Why is it gaining traction?

Hooks deliver deterministic enforcement—agents can't ignore or bypass them like prompts or skills—making it a reliable control layer over advisory measures. Devs get instant secret detection (AWS keys, GitHub tokens), PII blocking (SSNs, cards), and audit trails without setup hassle. The agent-trace compatibility lets you graph tool usage and failures, standing out in a sea of prompt-based "safety."

Who should use this?

AI-heavy devs in Cursor or Claude Code who generate code via agents and fear leaking keys or running rm -rf. Teams using OpenClaw for deterministic simulation testing or policy enforcement in agent workflows. Security-conscious coders evaluating deterministic security layers before production AI pipelines.

Verdict

Grab it if you're early on Cursor/Claude agents—37 stars and 1.0% credibility score signal raw potential, but it's WIP with regex limits, slow scans, and Cursor read-hook bugs. Solid docs and tests; productionize after fixes for a true deterministic security win.

Sign up to read the full AI review Sign Up Free

Similar repos coming soon.