antropos17

antropos17 / Aegis

Public

Independent AI Oversight Layer — monitors what AI agents do on your computer. EDR for AI Agents.

aegisprotect.vercel.app ai ai-agents ai-safety ai-security cybersecurity
59
14
100% credibility
Found Feb 22, 2026 at 9 stars 7x -- GitGems finds repos before they trend. Get early access to the next one.
Sign Up Free
AI Analysis
JavaScript
AI Summary

AEGIS is an open-source desktop app that monitors AI agents on Windows machines by tracking their processes, file accesses to sensitive areas, network connections, and behavioral anomalies via a real-time dashboard.

How It Works

1
🔍 Discover AEGIS

You hear about AI helpers on your computer that read files and connect online, and want a simple way to watch what they do.

2
📥 Download and Install

Grab the ready-to-run program for Windows from the latest release and double-click to set it up—no extra steps needed.

3
🚀 Launch Your Guardian

Open AEGIS and it quietly starts watching in the background, showing a cool radar dashboard right away.

4
🛡️ Spot AI Activity Live

Watch agents appear on the radar with risk colors, see file touches and connections pop up in real-time feeds, feeling in total control.

5
⚙️ Set Your Protection Level

Choose easy presets like Balanced or Strict to decide what to watch or block for each agent.

6
📊 Review Alerts and Reports

Check colorful timelines, agent cards, and optional smart summaries to understand any worries.

✅ Enjoy Peace of Mind

Your computer stays safe with ongoing oversight, custom logs, and alerts—AI agents behave while you stay informed.

Sign up to see the full architecture

5 more

Sign Up Free

Star Growth

See how this repo grew from 9 to 59 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 Aegis?

Aegis is a privacy-first Electron desktop app built with Svelte that runs on Windows, scanning for 95+ AI agents like GitHub Copilot, Cursor, or Devin via process signatures. It tracks their file accesses to sensitive spots (.ssh, .env, agent configs), outbound network connections with domain flagging, and behavioral anomalies against rolling session baselines—all locally with no telemetry. You get a real-time dashboard with radar viz, activity feeds, and exportable audit logs to oversee what these agents do on your machine.

Why is it gaining traction?

Unlike enterprise tools watching prompts sent to cloud AIs, Aegis flips it: real-time local monitoring of agent actions, with anomaly scores for file spikes or unknown endpoints. The neumorphic Svelte UI shines—animated radar orbits agents by risk, filterable feeds, and opt-in Anthropic analysis for threat reports. Prebuilt Windows exe means instant setup, no Node required.

Who should use this?

Developers wielding Cursor, Claude Code, or SWE-Agent who suspect sneaky file reads or phoning home. Security pros auditing autonomous agents like Devin before deployment. Anyone running AI IDE extensions or CLI tools, needing baselines to spot deviations without vendor lock-in.

Verdict

Solid alpha concept for AI oversight on Windows (10 stars, 1.0% credibility), but immature—lacks enforcement, Mac/Linux. Grab the exe if testing local agents; contribute to roadmap for cross-platform polish.

(198 words)

Sign up to read the full AI review Sign Up Free

Similar repos coming soon.