npx agentspec-cli install 5213447d-0081-4f01-9ef8-4bcf2e7c8a50 --file-type agents-md→ .agents/
# AGENTS.md ## Project Overview Kanmail is a desktop email client that functions like a kanban board, built with Wails v3 (Go backend + React frontend). It supports Mac, Windows, and Linux. ## Architecture Overview ### Backend (Go) **Entry Point**: `main.go` creates the Kanmail app and starts the Wails app. **Core Structure** (`internal/kanmail.go`): - `Kanmail` struct holds the Wails app and all services - Services are registered with Wails as bound Go methods callable from frontend - Circular dependencies are resolved via `Bootstrap()` pattern **Services** (`internal/services/`): - `AppService`: Application lifecycle, window management, license checking, analytics - `SettingsService`: User settings and configuration persistence - `AccountsService`: Manages email accounts (creates/deletes Account instances) - `EmailsService`: High-level email operations (send, fetch, etc.) - `ContactsService`: Contact management and avatars **Email Implementation** (`internal/emails/`): - `Account`: Represents a single email account with IMAP/SMTP connection pools - `Folder`: Represents an IMAP mailbox with UID tracking and pagination - Tracks UIDs from the last 90 days for pagination - Uses sparse UID → email cache populated during pagination - `lastSentUID` and `lastSentDate` track what frontend knows about - `ConnectionPool`: Manages pooled IMAP/SMTP connections (regular, priority, background) - `imapinterface/`: Abstraction layer allowing fake IMAP for testing/screenshots **Key Patterns**: - Connection pooling for IMAP/SMTP with priority levels - UID-based pagination for email folders (folders >1000 emails paginate) - Sparse caching with cache invalidation on UID validity changes - OAuth2 support with token refresh (see `emails/oauth/`) - Context-based logging with `zerolog` **Caches** (`internal/caches/`): - SQLite-based caching (avatars, contacts, email metadata, folder UIDs, licenses) - Single shared `caches.db` file managed by `Caches` struct **Types** (`internal/types/`): - `AccountSettings`, `FolderName`, `AccountName`, `Email` and other domain types - `AccountError` for wrapping account-specific errors - Event types for frontend communication ### Frontend (React + TypeScript) **Build System**: Vite with React plugin, TypeScript checking, Babel decorators support **Structure**: - `src/components/`: React components organized by feature - `emails/`: Email list, thread, message components - `settings/`: Settings UI components - `contacts/`, `send/`, `license/`, etc. - `src/stores/`: State management using custom store pattern - `base.tsx`: BaseStore class with subscription system - `emails/`, `settings.ts`, `contacts.ts`, etc. - Uses decorator pattern with `@subscribe` for component subscriptions - `src/util/`: Utility functions - `src/styles/`: LESS stylesheets - `src/wails/`: Auto-generated Wails runtime (DO NOT EDIT MANUALLY) - `src/bindings/`: Auto-generated Go service bindings (DO NOT EDIT MANUALLY) **State Management**: - Custom store pattern with `BaseStore` class - Components subscribe to stores via `subscribe()` decorator - Stores notify subscribed components on state changes ## Important Notes - NEVER edit files in `frontend/wails/` or `frontend/bindings/` - they are auto-generated - The project uses a custom fork of go-imap (`github.com/oxygem/go-imap/v2`) - see replace directive in `go.mod` - Email threading logic in `frontend/src/threading.js` - Connection pools maintain 2 regular + 2 priority + 1 background IMAP connections per account
Apple Mail PingZero MCP Server
An MCP server bridging AI agents and Apple Mail via AppleScript on macOS. **Stack:** Python 3.10+, FastMCP 4 / MCP SDK 2, AppleScript (via `osascript`) **Version:** v0.12.0 | **Tests:** 1812 unit / 44 e2e / 76 integration | **Coverage:** 90%
by AgentHub Crawler
AGENTS.md — agents-md-generator
`agents-md-generator` is a Python-based MCP (Model Context Protocol) server designed to generate and maintain `AGENTS.md` files—architectural context documents for AI coding agents. It uses Tree-sitter for high-fidelity AST analysis across multiple languages (Python, TypeScript, Go, C#) and `uv` for dependency management. The system operates as an incremental, cache-aware pipeline that transforms a codebase into a structured payload for AI assistants to perform maintenance tasks without manual e
by AgentHub Crawler
AGENTS.md
This is a LiveKit Agents project. LiveKit Agents is a Python SDK for building voice AI agents. See @README.md for more details on the application itself. The following is a guide for working with this project.
by AgentHub Crawler
Sign in to leave a comment.