AEGIS VSCode Extension

AI Agent Behavior Collector and Visualizer for VSCode.

Features

• File Monitoring: Automatically tracks file saves and changes
• Terminal Monitoring: Detects AI tool commands (claude, copilot, cline, codex)
• Event Timeline: Visual sidebar showing all captured events
• Prompt Version Management: Compare different prompt versions
• Evidence Export: Export complete event chains as JSONL
• AEF Protocol Integration: Uses AEF MCP for event normalization

Installation

Prerequisites

• VSCode 1.85.0 or higher
• Node.js 18+
• Python 3.8+ (for AEF MCP server)

Build from Source

# Install dependencies
npm install

# Install webview dependencies
cd webview
npm install
cd ..

# Compile TypeScript
npm run compile

# Package extension
npm run package

Usage

1. Install the extension in VSCode
2. The AEGIS sidebar will appear in the activity bar
3. Events are automatically captured as you work
4. Click on timeline events to view details
5. Use the toolbar to export evidence or verify chains

Configuration

The extension stores events in ~/.aegis/events.jsonl and prompts in ~/.aegis/prompts/.

Development

# Watch mode for development
npm run watch

# Run tests
npm test

# Lint code
npm run lint

Architecture

• src/extension.ts - Main extension entry point
• src/eventCollector.ts - File and terminal event monitoring
• src/aefClient.ts - AEF MCP integration
• src/timelineProvider.ts - Webview provider for timeline sidebar
• src/utils.ts - Utility functions
• webview/ - React-based UI components

License

MIT