Which coding agents does TeamAI CLI support?

TeamAI CLI's README lists Claude Code, Codex, Cursor, CodeBuddy, WorkBuddy, OpenCode, OpenClaw, Hermes, DeepSeek Harness, Qoder and ZCode. Coverage is not uniform: Claude Code, Codex, Cursor, CodeBuddy and Qoder tick every column in the support table, while OpenClaw, Hermes and DeepSeek Harness support a narrower set of resource types.

Where does TeamAI CLI store a team's shared skills and rules?

TeamAI CLI keeps everything in a git repository the team already owns. The README names GitHub, GitLab, GitCode, CNB, TGit and private git services as hosts. An admin creates the repo, grants write access to the team, and members run teamai init against its URL; after that every AI session pulls the latest published resources with no manual sync.

Can TeamAI CLI install resources for just one project?

Yes. Project scope is the default for TeamAI CLI: running teamai init inside a project directory installs the shared resources under that project. Passing --scope user instead installs them under the home directory so they apply to every project on the machine. Version 0.24.0-beta.2 also routes new project installs into per-project directories under ~/.teamai/projects.

What changed in TeamAI CLI v0.24.0-beta.2?

TeamAI CLI v0.24.0-beta.2 landed on September 9 with declarative team environment installs, first-class JoyCode and Qoder agent support, and a rename of the install command to packages. The same release reworked where machine data lives, fixed Codex token collection in the dashboard, and taught the contribution flow to recognise Japanese course-correction prompts.

Is TeamAI CLI free for commercial use?

Yes. Tencent publishes TeamAI CLI under the MIT license, which permits commercial use, modification and redistribution. The tool is distributed on npm as teamai-cli and the source lives at github.com/Tencent/teamai-cli, so a company can also fork it and run entirely against its own private git host.

Tencent · 2026-09-09 · major

TeamAI CLI — Tencent ships one AI setup for a whole team

TeamAI CLI is Tencent's MIT-licensed command-line tool that keeps a team's skills, rules, docs and MCP servers in one git repo and syncs them into Claude Code, Codex, Cursor, CodeBuddy, OpenCode and other agents. It passed 2,700 stars on September 9.

GitHub social preview for the Tencent teamai-cli repository

One git repo holds a team's agent skills, rules and MCP servers, and every member's coding agent pulls them automatically.

Quick facts

Maker	Tencent
License	MIT
Language	TypeScript
Install	npm install -g teamai-cli
Latest release	v0.24.0-beta.2, September 9, 2026
Latest stable	v0.23.1, September 9, 2026
GitHub stars	2,735

What is it?

TeamAI CLI turns a shared git repository into the source of truth for how a team's AI coding agents behave. An admin publishes skills, rules, docs, hooks, environment settings and MCP servers there; each member runs teamai init once, and from then on every AI session pulls whatever the admins have published. Tencent released it under MIT.

How does it work?

The tool splits into three layers. Team Execution covers init, pull and push plus the resources themselves, and is the part that is shipping today. Team Context, in beta, adds recall, shared learnings, a codebase graph and a team wiki so an agent understands the codebase. Team Improvement, also in beta, collects session analytics, usage digests and a dashboard.

Why does it matter?

Teams that adopt coding agents end up with the same rules copy-pasted into a dozen private config directories, drifting apart with every edit. Publishing them through a reviewed git workflow makes the setup auditable and gives new joiners a working configuration on day one. The multi-agent support also means the setup survives a member switching from Cursor to Claude Code.

Who is it for?

engineering leads standardising AI coding practice across a team

Frequently asked questions

Which coding agents does TeamAI CLI support?: TeamAI CLI's README lists Claude Code, Codex, Cursor, CodeBuddy, WorkBuddy, OpenCode, OpenClaw, Hermes, DeepSeek Harness, Qoder and ZCode. Coverage is not uniform: Claude Code, Codex, Cursor, CodeBuddy and Qoder tick every column in the support table, while OpenClaw, Hermes and DeepSeek Harness support a narrower set of resource types.
Where does TeamAI CLI store a team's shared skills and rules?: TeamAI CLI keeps everything in a git repository the team already owns. The README names GitHub, GitLab, GitCode, CNB, TGit and private git services as hosts. An admin creates the repo, grants write access to the team, and members run teamai init against its URL; after that every AI session pulls the latest published resources with no manual sync.
Can TeamAI CLI install resources for just one project?: Yes. Project scope is the default for TeamAI CLI: running teamai init inside a project directory installs the shared resources under that project. Passing --scope user instead installs them under the home directory so they apply to every project on the machine. Version 0.24.0-beta.2 also routes new project installs into per-project directories under ~/.teamai/projects.
What changed in TeamAI CLI v0.24.0-beta.2?: TeamAI CLI v0.24.0-beta.2 landed on September 9 with declarative team environment installs, first-class JoyCode and Qoder agent support, and a rename of the install command to packages. The same release reworked where machine data lives, fixed Codex token collection in the dashboard, and taught the contribution flow to recognise Japanese course-correction prompts.
Is TeamAI CLI free for commercial use?: Yes. Tencent publishes TeamAI CLI under the MIT license, which permits commercial use, modification and redistribution. The tool is distributed on npm as teamai-cli and the source lives at github.com/Tencent/teamai-cli, so a company can also fork it and run entirely against its own private git host.

Try it

npm install -g teamai-cli && teamai init https://github.com/yourorg/yourrepo