AI Coach is a VS Code / Cursor extension that gives you a single sidebar panel to browse AI skills, manage MCP servers, audit your AI tools, and track usage — all powered by coach.omazy.ai.
Skills are also distributed on the Skillfish network (npx skillfish add ...).
Features
Explore — Browse & Install
Search and browse skills and MCP servers from coach.omazy.ai. Each card shows name, description, tags, and upvote count. One-click install writes SKILL.md to your Cursor skills directory or merges MCP config into mcp.json.
Installed — Manage What's Active
See all installed skills and MCP servers, split by scope (global ~/.cursor/ vs project .cursor/). Uninstall skills, open config files, and toggle MCP servers.
My Setup — AI Environment Audit
Auto-detects AI tools installed on your system:
| Category |
Tools Detected |
| Editors & IDEs |
Cursor, VS Code, Windsurf |
| CLI Tools |
Claude Code, GitHub Copilot CLI, Aider, Gemini CLI, OpenAI Codex CLI |
| Extensions |
Continue.dev, Tabnine |
Also scans your workspace for AI context files (CLAUDE.md, .cursorrules, .windsurfrules, .github/copilot-instructions.md) and shows an MCP compatibility matrix.
Usage — Token & Cost Tracking
Reads local usage data from Claude Code (~/.claude/) and Cursor (~/.cursor/) when available. Shows token counts, request counts, cost estimates, and model breakdown. Status bar items update every 60 seconds.
Getting Started
- Install the extension from
.vsix or VS Code Marketplace
- Click the AI Coach icon in the activity bar (left sidebar)
- Browse the Explore tab and install your first skill
Configuration
| Setting |
Default |
Description |
aicoach.apiBaseUrl |
https://coach.omazy.ai |
API origin for skills and MCP data |
aicoach.skills.installLocation |
user |
Install to ~/.cursor/skills/ or workspace |
aicoach.mcp.configTarget |
user |
Write MCP config to ~/.cursor/mcp.json or workspace |
cursorUsage.statusBar.enabled |
true |
Show usage in the status bar |
Architecture
Built on the CIR (Component-Interface-Registry) pattern. Each sidebar tab is a self-contained feature that implements FeatureProvider and registers with a central FeatureRegistry. Adding a new tab is one file + one line.
See docs/ARCHITECTURE.md for details.
Development
cd packages/vscode-aicoach
pnpm install
pnpm run compile # build once
pnpm run watch # dev mode
pnpm run package # produce .vsix
Press F5 in VS Code / Cursor to launch the Extension Development Host.
See docs/DEVELOPMENT.md for the full development guide.
Links