Aegis AI — Secure Access Layer
Manage, govern, and audit your AI agents - all from the VS Code sidebar.
Features🔌 50+ MCP Servers — Browse and install from a built-in catalog: GitHub, Playwright, PostgreSQL, Slack, Stripe, MongoDB, and many more. One click to add, one click to go. 🛡️ Policy Engine — Control what your AI can do. Set 🔐 Encrypted Secrets — Store API keys and tokens with encryption. Secrets are scoped per server and injected automatically — no plaintext credentials in config files. 📋 Audit Log — Every tool call is logged: timestamp, server, tool name, result, and duration. Query, filter, and export as JSON or CSV. 🚀 Gateway Proxy — A single MCP endpoint that routes to all your servers. Tool namespacing prevents conflicts. Policies are enforced on every call. 🔄 VS Code Sync — Auto-generates Getting Started
Data StorageAll data is stored locally in Requirements
LinksFAQQ: Where is my data stored?
A: All data lives locally in Q: Does it work with agents other than VS Code Copilot?
A: Yes. Any MCP-compatible client (Claude Desktop, Cursor, Windsurf, etc.) can connect to the Aegis gateway at Q: Can I add my own MCP servers not in the catalog? A: Yes. Use Add MCP Server → choose npm package, custom command, or remote HTTP endpoint. Q: How are secrets protected? A: Secrets are encrypted using a 256-bit random master key stored with restricted file permissions. The encrypted file is useless without the key. Q: What happens if I deny a tool in a policy? A: The AI agent receives a "denied by policy" error. The attempt is logged in the audit trail. Q: How do I reset everything?
A: Delete the Feedback & SupportFor issues, feature requests, or licensing inquiries: officialumangbarthwal@gmail.com LicenseLicensed under the Apache License, Version 2.0. Copyright 2026 Umang Barthwal. |
