Local flight recorder and execution boundary for AI coding agents. It records every tool call, blocks the risky ones until you approve, and shows it all in a local cockpit. Works with Cursor, Claude C
View on GitHub ↗Homepage ↗The local flight recorder on the stdio wire for MCP & custom agents.
Website & Live Docs · Setup Playbook for AI Agents · Architecture Blueprint · Releases
Agent Sentinel is a single self-contained Go binary that sits on the stdio wire between an AI coding agent and your operating system. It records what the agent does, measures execution latency in milliseconds, halts dangerous commands before they touch your filesystem, and renders the entire session in a local web cockpit.
- 100% Loopback & Zero Cloud: Runs entirely on 127.0.0.1. No telemetry leaks, no cloud accounts, no API proxy intermediaries. - Tested & Verified Live Integrations: Natively supervises Claude Code and Cursor Agent via their respective hook lifecycle APIs, as well as any MCP server as a stdio proxy. - Single Static Executable: Compiled with the React + Tailwind web cockpit embedded (go:embed). No Node.js or background daemon required in production.
The Sentinel Cockpit: Real-time event timeline, token/latency telemetries, visual diff viewer, and human approval dock.
Capability Category Description :--- :--- :--- 🔍 Raw JSON-RPC Inspector Wire Level Every tools/call, tools/list, notification, request payload, and response is intercepted and displayed in real time in the cockpit. No guesswork. ⏱️ Latency & Traffic Telemetry Performance Track precise millisecond execution times per tool call (p50 / p95). Spot lagging servers, token burn rate, and runaway loops before they drain machine resources. 📼 JSONL Flight Recorder & Replay Offline Replay Every session is saved locally to ~/.sentinel/sessions/.jso
From the project README.
Add the radar badge to your README — it shows your project was picked up by MCP Radar and links to this page:
[](https://mcp.liqiwa.com/s/zaninifrancesco--agent-sentinel.html)
Git-native persistent memory for AI coding agents. Implements Google OKF v0.2 with sub-300µs in-memory BM25 search, embedded MCP server, and progressive disclosure. Slashes token bloat by 80% with zer
vshulcz/deja-vuYour agents already solved this. deja finds it — retroactive search and MCP recall over the session logs your coding agents already wrote. Cross-agent, SSH-portable, one zero-dep binary, fully local.
olucurious/watchtowerSelf-hosted error tracking. A drop-in for Sentry and AppSignal SDKs: one Go binary and Postgres, with Slack, Linear, email and an MCP server for coding agents.
KnifeLemon/DroidlineAndroid automation from code. Test your apps on real phones, script the tasks you repeat, run a shelf of phones, or let an AI agent use one, from Python, Node.js, the CLI, HTTP or Appium clients. No A
enowdev/succubusCross-agent coordination for AI coding agents. One daemon, one database — so multiple agents in one repo can see each other's plan, tasks, and file claims.
AsWali/CallBoardA shared to-do list for you and your AI coding agents
The top new MCP servers of the week, every Monday. No spam, unsubscribe anytime.