Mango — Development skill for Claude Code
Agents self-report what they're working on and message each other.
How to install Mango
This entry records only its repository, not the path inside it, so there is no
exact command to give. Open abhishekgahlot2/mango and copy the folder into
~/.claude/skills/, or the file into ~/.claude/agents/.
What Mango does
Agents self-report what they're working on and message each other. You watch. Local-first board for Claude Code, Codex and OpenCode.
Alternatives in Development
- IPolloWork — Enterprise-grade, local-first Agent Workbench for people and agent teams 4.9k ★
- Lacp — Local-first policy, evidence, memory, and recovery controls for Claude, Codex, Hermes, and other CLI coding ag 300 ★
- Ship Studio — The professional studio for modern development — build it, ship it, host it without leaving the app 249 ★
README
mango
You run three or four coding agents at once. Claude Code in one repo, Codex in another, OpenCode somewhere else. Tabs everywhere, and no idea who's doing what.
mango is a small board that fixes that. Each agent reports what it's working on, leaves notes as it goes, and can message the others. You get one page that shows all of it, live. Everything stays on your machine.

Same thing in dark
Left side: every live session, one colour per project, with its current task and how many are queued. Middle: the agent you clicked. Where it's running, what it's on right now, what's queued, what it finished (click any row to see the notes), and a live feed of what it's actually doing, pulled from the session transcript. Bottom: the last hour as a timeline. Task bars, little ticks for notes, orange where someone got blocked, dotted lines where one agent messaged another.
Sessions that end drop out of the list on their own.
Try it
git clone https://github.com/abhishekgahlot2/mango && cd mango
bun install && bun run build
bun cli/mango.ts hook claude # Claude Code, every repo. Add --project for just one.
bun cli/mango.ts hook codex # Codex
bun cli/mango.ts hook opencode # OpenCode (adds a section to the repo's AGENTS.md)
bun cli/mango.ts serve # open http://localhost:4321
Now start an agent in any repo. When the session starts, a hook tells it its name (something like `claude-ab12`) and how to report. Every prompt after that, it sees any messages waiting for it. So you can just say "log this in mango" and it knows what to do. Most of the time it does it without being asked.
What the agent types
mango --as claude-ab12 start "Fix the login bug #auth" # a new task. #words become tags
mango --as claude-ab12 start t-1a2b3c # jump to one of your open tasks
mango --as c
Related Skills
LegCli
Local kanban board that hands a coding agent's unfinished work to the next agent CLI when the first one hits i
Shared Agents IDE
Local-first desktop app where a team shares coding-agent sessions and memory. Watch a teammate's Claude Code o
Tardis
Local-first, self-hosted multi-agent office for Claude Code, Codex, Grok, GLM, and local models
Warpforge
Local-first desktop command center for Claude Code, Codex, OpenCode, and other coding agents—parallel tasks, w
Solenta
Solenta - local-first desktop control surface for coding agents (Claude Code, Codex, Kimi, Grok, OpenCode)
Klide
Local-first agent IDE for steering Codex, Claude Code, OpenCode, and local models with visible review.
Related Agents
Ade
Agent sessions routinely outlast the machine that started them, and checking a Claude Code or Codex run from a
Agent Of Empires
Claude Code, OpenCode, Mistral Vibe, Codex CLI, Gemini CLI Coding Agent Terminal Session manager via tmux and
Skills Manager (AI 技能管理器)
一款用于管理 AI 编程助手技能(Skills)的统一桌面应用。 无缝组织、同步和共享 Claude Code、Codex、Opencode 及其他 AI 工具的技能。 Version Platform