Tencent/teamai-cli

What is TeamAI‑CLI?

TeamAI‑CLI is a command‑line tool (published on npm) that lets a software team treat its AI‑related assets – prompts, custom agents, rules, environment configs, model references, and even usage logs – as first‑class code stored in a shared Git repository. Once a single teammate runs teamai init <repo‑url> the tool automatically syncs the latest “team AI” files to every member’s local workspace, so every AI session (e.g., in Claude, Codex, Cursor, etc.) works with the same set of skills, policies, and models.

Why it matters

Consistency – All team members get the same prompt libraries, safety rules, and model versions without manual copying.
Version‑controlled AI assets – Skills, hooks, and even model references live in Git, so you can review, branch, and roll back AI changes just like code.
Zero‑touch updates – After the initial setup, the CLI pulls new assets on every AI interaction, keeping everyone in sync.
Works with many AI agents – The README lists dozens of supported agents (Claude, Codex, Cursor, GitHub Copilot CLI, etc.), each automatically receiving the team’s configuration.

Quick start (one‑liner for your AI assistant)

Install the teamai skill: https://github.com/Tencent/teamai-cli/tree/main/skills/teamai , load the teamai skill, then set up TeamAI for my team from scratch.

After the assistant runs the above, you can manage the team with simple chat commands, e.g.:

/teamai Help me set up TeamAI for my team from scratch
/teamai Help me join my team's TeamAI, repo URL is https://github.com/your-org/your-repo
/teamai Share my xxx skill with the team
/teamai Open the TeamAI dashboard

Installation

# Global install via npm
npm install -g teamai-cli

Core workflow

Create a shared Git repo (GitHub, GitLab, GitCode, CNB, TGit, or any private Git service). Protect the default branch and give teammates push rights.
Initialize – Run teamai init <repo‑url> either in a workspace directory (default) or with --scope user to install under ~/.
Team admin can start from a ready‑made template in the teamai‑hub org and fork it.
Team members simply run the same teamai init command; the CLI clones the repo and sets up local hooks.
Every AI session now automatically pulls the newest skills, rules, environment definitions, model references, and other assets.

Main capabilities (layered on Git)

Layer	What it does
Team Execution	Distributes skills, rules, environment variables, agents, hooks, MCP servers, and model references to every teammate’s AI tool.
Team Context (beta)	Provides agents with shared learnings, a code‑base graph, and a team wiki so the model can answer with team‑specific knowledge.
Team Improvement (beta)	Collects usage metrics and session logs to help the team understand how AI is being used and where to improve.

Supported agents (excerpt)

Agent	Skills	Rules	Env	Agents	Hooks	MCP	Models
Claude Code	✅	✅	✅	✅	✅	✅	✅
Codex	✅	✅¹	✅	✅	✅	✅	✅
Cursor	✅	✅	✅	✅	✅	✅	—
GitHub Copilot CLI	✅	✅	✅	✅	✅	✅	—
… (many more)

*¹ Rules stay on regardless of file path. *² Environment variables are applied only via teamai env exec.

Documentation & learning resources

Getting Started – docs/guide/getting-started.md
Usage Guide – daily workflows and command reference (docs/usage-guide.md)
Product Overview – architecture and distribution controls (docs/product-overview.md)
Git Providers – supported hosting services (docs/guide/providers.md)
Windows Setup – special notes for hooks and shells (docs/guide/windows.md)
Technical Designs – deeper design docs (docs/designs/README.md)

Who maintains it?

The project is open‑source under the MIT license, authored by Tencent. Contributions are welcomed via the usual GitHub workflow (issues, PRs, and a CONTRIBUTING.md guide). A contributors badge shows community activity.

TL;DR

TeamAI‑CLI turns a Git repo into a single source of truth for all AI assets a team uses. Install globally, point it at a shared repo, and every teammate’s AI tools automatically stay in sync with the team’s prompts, policies, models, and usage analytics.