← Back to Skills Marketplace
203
Downloads
0
Stars
1
Active Installs
5
Versions
Install in OpenClaw
/install team-collab
Description
多模型团队协作系统 - 像公司一样运作,支持三个专业团队:软件开发团队(8人)、搞钱特战队(10人)、小说+漫剧团队(10人)。固定子窗口模式,每个角色分配最优模型。触发词:"软件开发团队"、"搞钱特战队"、"小说团队"。
Usage Guidance
This skill appears coherent and implements what it claims: role-based model assignments, spawn-call examples, templates, and simple local project scaffolding. Before installing: 1) Confirm your OpenClaw/platform account already provides access to the model IDs referenced (bailian/qwen-*, deepseek-coder); if those models require provider credentials, ensure you configure them through the platform rather than embedding secrets. 2) Inspect the package repository (package.json points to a GitHub URL) if you plan to run the scripts locally. 3) The included scripts are benign (create folders, write JSON, print spawn examples) — avoid running any code you do not trust in environments that contain sensitive data. 4) If you expect the skill to spawn models autonomously, confirm platform permissions and quotas—autonomous model spawning is normal but increases resource/attack surface. Overall, no hidden network calls, credential access, or obfuscated code was found in these files.
Capability Analysis
Type: OpenClaw Skill
Name: team-collab
Version: 3.1.0
The skill bundle provides a comprehensive framework for multi-model team collaboration, organizing AI agents into specialized roles for software development, business strategy, and content creation. It includes Python scripts (`start-project.py` and `team-workflow.py`) for project initialization and generating workflow instructions, along with detailed configuration files and templates. No evidence of malicious intent, data exfiltration, or harmful prompt injection was found; the bundle's behavior is consistent with its stated purpose of managing multi-agent workflows.
Capability Assessment
Purpose & Capability
Name/description (multi-model team collaboration) match the SKILL.md, config, and scripts. The package.json, config/default.json, references, and scripts all implement role/model assignment, spawn-call examples, templates and project scaffolding consistent with the stated purpose. No unrelated services or credentials are requested.
Instruction Scope
SKILL.md instructs organizing roles, using triggers, and spawning model-specific sub-windows; the included scripts only print spawn instructions or create local project scaffolding. There are no instructions to read arbitrary system files, exfiltrate data, contact unknown endpoints, or access environment variables beyond what the platform would normally provide.
Install Mechanism
No install spec is provided (instruction-only style with a couple of utility scripts). The files are static and the scripts perform local I/O only (directory creation and printing). No downloads from external URLs or archive extraction are present.
Credentials
The skill declares no required env vars or credentials, which is consistent with the code (no direct network/curl/auth usage). It assumes the hosting platform supplies model/session APIs (the scripts show sessions_spawn usage as examples). If model providers used (e.g., bailian/qwen-*) require separate credentials on your platform, that is an external concern but not an inconsistency in the skill itself.
Persistence & Privilege
always is false and the skill does not request system-level persistence or modify other skills. The README references optional config file locations (~/.openclaw/team-collab.json and ./team-collab.json) which are reasonable for user configuration and not privileged by themselves.
How to Use
- Make sure OpenClaw is installed (local or Docker)
- Run the install command in chat:
/install team-collab - After installation, invoke the skill by name or use
/team-collab - Provide required inputs per the skill's parameter spec and get structured output
Version History
v3.1.0
Team Collab v3.1.0 introduces "fixed sub-window mode" and optimal model assignment per role.
- 新增“固定子窗口模式”:每个角色独立窗口,拥有记忆,不重复新建
- 每个团队成员分配最优模型(qwen3.5-plus / glm-5 / qwen-plus)
- 子窗口标签命名规范化,便于快速定位
- 质量把控员职责加入“修bug”
- 文档整体结构优化,突出模型分配和窗口管理方式
v3.0.0
**Team-collab 3.0.0 introduces multi-disciplinary, specialized teams beyond software development.**
- 新增“小说+漫剧团队”,支持 AI 小说、漫剧创作,10人全职能配置
- “软件团队”更名为“软件开发团队”,描述/触发词同步调整
- 三大专业团队可并行调用:软件开发(8人)、搞钱特战队(10人)、小说+漫剧团队(10人)
- 明确各团队专属角色/分工、项目流程与方法论
- 快速触发词&自定义角色/模型机制同步升级
- 更新文档结构/表格,更便于理解与场景适配
v1.1.1
team-collab v1.1.1
- Added initial configuration files: config/README.md, config/default.json, and package.json.
- Updated SKILL.md with refined descriptions and/or instructions.
- Improved and updated scripts/team-workflow.py.
- Lays groundwork for custom configuration and easier setup.
- Minor documentation and workflow clarifications.
v1.1.0
新增高级模式:真正切换模型协作
v1.0.0
多模型团队协作系统 - 像公司一样运作
Metadata
Frequently Asked Questions
What is Team Collab?
多模型团队协作系统 - 像公司一样运作,支持三个专业团队:软件开发团队(8人)、搞钱特战队(10人)、小说+漫剧团队(10人)。固定子窗口模式,每个角色分配最优模型。触发词:"软件开发团队"、"搞钱特战队"、"小说团队"。 It is an AI Agent Skill for Claude Code / OpenClaw, with 203 downloads so far.
How do I install Team Collab?
Run "/install team-collab" in the OpenClaw or Claude Code chat to install it in one step — no extra setup required.
Is Team Collab free?
Yes, Team Collab is completely free, licensed under MIT-0. You can download, install and use it at no cost.
Which platforms does Team Collab support?
Team Collab is cross-platform and runs anywhere OpenClaw / Claude Code is available (cross-platform).
Who created Team Collab?
It is built and maintained by davidme6 (@davidme6); the current version is v3.1.0.
More Skills