- Published on
Tencent Cloud Open-Sources Internal TeamAI Tool, Enabling Cross-Agent Skill Sharing
- Authors
- Name
- DeepHorus
On October 10, Tencent Cloud announced the open-sourcing of TeamAI, which it had been using internally. This is a Git-based team AI collaboration tool. It puts team Skills, work specifications, tool configurations, and project knowledge into a Git repository, then syncs them to members' Agents, achieving "configure once, reuse by everyone."
TeamAI manages team AI assets the way code is managed, supporting review, updates, and version management. Currently, it has been adapted for 16 kinds of Agents, including WorkBuddy, CodeBuddy, Claude Code, Codex, and Cursor. Its goal is to reduce repetitive configuration and make team experience reusable.
In terms of the collaboration workflow, TeamAI follows the Git approach. When a member submits a skill or modifies a specification, the system automatically creates a branch and initiates a merge request, which is handed over to the team for review. After it is merged, Agents that support a session-start Hook will pull updates at the beginning of a new conversation, with no need to notify each person individually or replace things manually.
Recently, TeamAI added the capability to uniformly distribute model configurations. For tasks involving internal data and knowledge, administrators can uniformly configure model services that meet the team's security requirements. This can reduce the risk of members connecting on their own and choosing the wrong model.
TeamAI will also, based on situations such as repeated corrections and failed tool calls, prompt members to share their experience at the end of a session. The Agent organizes the problem and the solution into a document and stores it in the team repository. The next time it handles a task, the Agent can retrieve existing experience and materials as a reference for solving it.
After installing the TeamAI Skill, administrators and members can operate it in natural language within already-adapted AI tools. TeamAI has been integrated into Tencent Cloud's agent management platform ClawPro. Members can connect local Agents to ClawPro, and administrators can uniformly manage configurations such as Skills and MCP, and distribute them to connected Agents.
From the perspective of functionality, TeamAI targets problems in team AI usage such as scattered configuration, experience that is hard to accumulate, and inconsistent model access. Through the Git workflow and ClawPro management and control, teams can reduce the cost of repetitive configuration and notifying people one by one. Currently, the tool has been adapted for 16 kinds of Agents and has entered the ClawPro management and control system.
