Usylli

Lagless — AI skill for Claude Code

AI community

A native macOS task manager for the age of AI agents: see every Claude Code session, Codex thread and MCP server, why each process runs, and free up disk space safely.

How to install Lagless

This entry records only its repository, not the path inside it, so there is no exact command to give. Open Usylli/lagless and copy the folder into ~/.claude/skills/, or the file into ~/.claude/agents/.

What Lagless does

A native macOS task manager for the age of AI agents: see every Claude Code session, Codex thread and MCP server, why each process runs, and free up disk space safely.

Alternatives in AI

Codeburn — Free, local tool to track AI coding token usage and cost across 37 tools and agents (Claude Code, Cursor, Code 9.7k ★
Nodeterm — Node-based terminal manager for AI coding agents — tmux-backed terminals and parallel agent sessions as dragga 1.3k ★
Agentrq — AgentRQ: Human-in-loop realtime conversational task manager for AI Agents 1.1k ★

README

Lagless

A native macOS task manager for the age of AI agents.
See what is running, why it is running and what is slowing your Mac down —
every Claude Code session, Codex thread and MCP server included.

Lagless showing AI agent sessions: what each is doing, its last prompt and reply, and the processes it launched

Why Lagless?

Activity Monitor shows hundreds of processes named `node`, `zsh` and `Helper (Renderer)` and leaves you to guess. Meanwhile coding agents quietly start MCP servers, dev servers, test runners and nested agents — and keep them alive for days after you stopped looking.

Lagless groups every process by what it belongs to, explains why it is running, and shows the agent behind it: what the session is about, what it was last asked, whether it is waiting for you, and everything it launched. When the Mac lags, it tells you why in one sentence and offers a safe fix.

Features

🤖 Agents

Claude Code sessions with their state (working, needs permission, waiting for input, idle) and for how long, last prompt and reply, model, context size, permission mode, git branch, tools used, edited files and pull requests.
*Codex