⏳ This skill is pending AI review.

Scores will appear once the review pipeline completes.

v1.0.0

agent-team-manager

countbot-ai@countbot-ai⭐ 785 stars· last commit 16d ago· 14 open issues

多智能体团队管理。创建、查看、修改、删除 CountBot 的多智能体团队,管理团队成员(角色)和团队级自定义模型配置。当用户要新建 Pipeline/Graph/Council 团队、调整成员分工、修改依赖关系、开关技能系统、设置团队专属模型时使用。

Choose how to use this skill

You do not need every option. Choose the path your AI client supports. The stable page stays the same; versioned files are immutable.

1. Native installer

This listing has no registered native installer command. Use the complete package or source fallback below, depending on what your client supports.

Do not guess an installer command or replace an existing version without reviewing the diff.

2. Complete package recommended

Download the ZIP when available. It includes SKILL.md plus the references, security notes and version metadata.

No complete ProSkills package is published for this listing yet.

3. Prompt-only

Copy the prompt above when the agent can read the stable page or when you want to adopt the workflow without installing a skill.

Need only the instruction file?

Download SKILL.md only if your client requires a single file. The complete ZIP is safer for a full installation because it preserves the references and release context.

No path installs or executes anything by itself. Your agent still needs access to the project files. Before updating, compare the installed version and review the diff.

—/10

// RATINGS

⭐GitHub Stars
⭐⭐⭐⭐ 785 on GitHubGitHub ↗

Popular

🟢ProSkills Score
—
📍

Not yet listed on ClawHub or SkillsMP

// README


What Is CountBot

CountBot is a lightweight, low-footprint AI assistant that runs on your server and stays up. A modest machine is enough to host it. Once it is bound to the chat tools you already use, you can call it directly from Lark, WeCom, WeChat or QQ, or work with it through the web interface.

Unlike a plain chat bot, CountBot can actually execute work. It reads and writes files, runs commands, searches the web, takes screenshots and sends media. Beyond that, you can wrap your recurring routines into skills, let it run on a schedule, and have the results delivered to the channel you choose.

Key points below.

  • Entry points are the IM tools you already use, including Lark, WeCom, WeChat, QQ, DingTalk, Telegram and Weibo, plus ServerChan for push-only delivery
  • Recurring routines can be wrapped into skills, triggered by a single instruction or run automatically by the scheduler
  • Scheduled tasks combine with proactive greetings to push digests, reminders and summaries on plan
  • Long-term memory keeps context across sessions, so earlier instructions do not have to be repeated
  • 22 model providers are built in, and models can be assigned per session or per task

Skills

The skill system is the core mechanism of CountBot.

A lot of daily work follows fixed procedures. Compiling a competitive digest every week, summarizing new content from a few sites every day, cleaning spreadsheets by a fixed rule. You have done these many times and know the steps by heart, yet still run them by hand each time.

Skills fix that. Write the procedure once, and afterwards a single instruction triggers it, or the scheduler runs it automatically.

Writing one is not hard. Describe when to use it and how it works, and that is enough. If you are not sure where to start, the built-in skill-creator walks you through the whole process, from clarifying the goal to validating the result. Finished skills can be enabled, disabled and configured from the panel, or installed from the SkillsHub.


Multi-Agent Teams

For more complex work, a single agent tends to lose track of things. Agent Teams let you group several roles, each owning one part, working in stages or in parallel.

Three orchestration modes are available.

pipeline runs stages in sequence, where the output of one stage becomes the input of the next. It suits tasks with a clear order of steps.

graph lets you declare dependencies between roles first, then runs everything without unmet dependencies in parallel. It suits tasks with branches and a final merge. If the dependencies form a cycle, the system reports an error instead of looping.

council has each member produce an independent result from their own perspective, then review each other, and finally consolidate. It suits judgment calls that need several angles weighed.

Once a team exists, send /team <team name> <task> in a session to invoke it, or mention the team name with @.

A team can carry its own model configuration. When configured, execution inside the team uses the team configuration; otherwise it inherits the model of the current session.


Session-Level Model Configuration

Changing models used to mean editing the global default, which affected every session. Now it can be switched inside a session.

A model picker sits next to the chat input. The selection applies only to the current session and leaves the global default untouched. Right beside it, the reasoning effort for that session can be adjusted, with the available levels determined by the model.

Switching the level on the same model is enough to cover different scenarios. Long tasks get a higher level, quick questions drop back to a lower one, without repeated trips to the settings page.

Order of precedence is as follows. The model selected in a session applies to that session; if a team has its own model configured, execution inside the team follows the team configuration; when neither is configured, the global default applies.


What It Can Do For You

ScenarioDescription
Daily digest deliveryScheduled tasks fetch and summarize the topics you specify by cron, then push to Lark, WeChat, QQ, ServerChan and other channels
In-group assistantMention the bot in a group to ask questions, look up material, draft content and call tools
Scheduled tasks and remindersRun arbitrary instructions on a schedule, with a selectable delivery target and bot account, and a selectable model per task
Proactive greetingsA heartbeat mechanism starts conversations or reminders at the frequency you set, suited to schedule reminders and shift handover briefings
Material handlingWeb scraping, document parsing, knowledge base retrieval, file read/write and batch processing
Multi-role collaborationSplit complex work across an Agent Team, run it in parts and consolidate
External coding toolsConnect Claude Code, Codex, OpenCode and dsh (DeepSeek Harness), either as callable tools or as agents bound to a channel

Supported Channels

ChannelDirectionNotes
LarkTwo-wayEvent subscription, supports groups and direct messages
WeComTwo-wayEnterprise internal bot
WeChatTwo-wayMultiple account binding supported
QQTwo-wayQQ Open Platform bot
DingTalkTwo-wayEnterprise internal bot
TelegramTwo-wayDirect Bot API
WeiboTwo-wayWeibo bot
XiaozhiTwo-wayXiaozhi AI hardware integration
ServerChanPush onlyA single HTTP request delivers a notification to WeChat, with no persistent connection, suited to pure broadcasting

All two-way channels share the same session model. A session processes one turn at a time, so concurrent messages from multiple entry points do not interfere with each other.


Supported Models

22 providers are built in, covering mainstream Chinese and international models, along with OpenAI-compatible custom endpoints and local inference.

openrouter   anthropic   openai      deepseek   moonshot   zhipu
groq         mistral     cohere      together   qwen       hunyuan
ernie        doubao      yi          baichuan   minimax    vllm
ollama       lm_studio   custom_openai          custom_anthropic

The request protocol is determined by the provider setting or by user selection.

ProtocolEndpoint
OpenAI Chat Completions/chat/completions
OpenAI Responses/responses
Anthropic Messages/v1/messages

A single provider can hold multiple models, fetched from the upstream /models endpoint or entered manually. Each model records its context window, image input capability, reasoning levels and related metadata. The global default, session level and scheduled task level can each designate which model to use.


Core Capabilities

ModuleDescription
Skill systemTurn recurring routines into skills that load automatically as callable capability. Supports enable/disable, parameter configuration and SkillsHub installs, with skill-creator built in for guided creation
Multi-agent teamsThree orchestration modes, pipeline, graph and council, invoked through /team or @, with optional team-level model configuration

// HOW IT'S BUILT

// FILES

2 files shipped with this skill. Inline text is the catalog copy. The upstream link is the source repository.

  • workspace/skills/agent-team-manager/SKILL.md10 KBupstream ↗
    ---
    name: agent-team-manager
    description: 多智能体团队管理。创建、查看、修改、删除 CountBot 的多智能体团队,管理团队成员(角色)和团队级自定义模型配置。当用户要新建 Pipeline/Graph/Council 团队、调整成员分工、修改依赖关系、开关技能系统、设置团队专属模型时使用。
    version: 1.0.0
    always: false
    ---
    
    # 多智能体团队管理
    
    通过命令行管理 CountBot 的多智能体团队,覆盖团队 CRUD、成员 CRUD、以及团队级模型配置。
    
    ## 使用场景
    
    - 用户说“帮我新建一个多智能体团队” -> 创建团队
    - 用户说“做一个文档深度分析团队” -> 按模板创建团队
    - 用户说“把这个团队改成依赖图模式” -> 修改团队
    - 用户说“给团队加一个审稿角色” -> 添加成员
    - 用户说“把 analyzer 的任务改一下” -> 修改成员
    - 用户说“把这个角色的提示词优化一下” -> 先看当前成员配置,再修改 `task` 和/或 `system_prompt`
    - 用户说“删掉 summarizer 这个角色” -> 删除成员
    - 用户说“给这个团队单独配置模型” -> 配置团队自定义模型
    - 用户说“看看有哪些团队/成员” -> 列表或详情
    
    ## 调用方式
    
    所有操作通过 `exec` 工具执行:
    
    ```bash
    python3 skills/agent-team-manager/scripts/agent_team_manager.py <command> [args]
    ```
    
    ## 严格语法
    
    - 只有这些一级命令:`list`、`info`、`template-list`、`create`、`update`、`delete`、`member-list`、`member-add`、`member-update`、`member-delete`、`config`、`config-set`、`config-reset`
    - `info` / `member-list` / `update` / `config-set` 的团队参数都是位置参数,不支持 `--team`
    - `member-update` 的成员参数是第二个位置参数 `member_ref`,不支持 `--id`
    - 开关技能系统要用团队命令 `update "团队名" --enable-skills` 或 `--disable-skills`
    - 先看 `--help`,再按帮助里的位置参数顺序执行;不要自行发明子命令或参数名
    
    ## 常用命令
    
    ### 列出团队
    
    ```bash
    python3 skills/agent-team-manager/scripts/agent_team_manager.py list
    ```
    
    ### 查看团队详情
    
    `team_ref` 支持团队名称、完整 ID、ID 前缀。
    
    ```bash
    python3 skills/agent-team-manager/scripts/agent_team_manager.py info "文档深度分析"
    ```
    
    ### 创建团队
    
    ```bash
    # 创建空团队,后续再逐个添加成员
    python3 skills/agent-team-manager/scripts/agent_team_manager.py create \
      --name "文档深度分析" \
      --description "理解文档 → 提取要点 → 分析问题 → 生成总结报告" \
      --mode pipeline \
      --enable-skills
    
    # 直接按内置模板创建
    python3 skills/agent-team-manager/scripts/agent_team_manager.py create \
      --template document-analysis
    ```
    
    ### 修改团队
    
    ```bash
    python3 skills/agent-team-manager/scripts/agent_team_manager.py update "文档深度分析" \
      --mode graph \
      --description "先并行抽取,再汇总结论" \
      --active
    
    python3 skills/agent-team-manager/scripts/agent_team_manager.py update "文档深度分析" \
      --disable-skills
    ```
    
    ### 删除团队
    
    ```bash
    python3 skills/agent-team-manager/scripts/agent_team_manager.py delete "文档深度分析"
    ```
    
    ## 成员管理
    
    ### 列出成员
    
    ```bash
    python3 skills/agent-team-manager/scripts/agent_team_manager.py member-list "文档深度分析"
    ```
    
    ### 添加成员
    
    ```bash
    python3 skills/agent-team-manager/scripts/agent_team_manager.py member-add "文档深度分析" \
      --id reader \
      --role "文档理解专家" \
      --task "通读文档,理解整体结构和核心内容" \
      --system-prompt "你是文档理解专家,先识别文档结构,再提炼核心主题。"
    ```
    
    Graph 模式可附带依赖与条件:
    
    ```bash
    python3 skills/agent-team-manager/scripts/agent_team_manager.py member-add "代码检查团队" \
      --id refactor \
      --role "重构建议专家" \
      --task "基于前序检查结果给出重构建议" \
      --depends-on syntax-checker,logic-analyzer \
      --condition-type output_contains \
      --condition-node logic-analyzer \
      --condition-text 严重
    ```
    
    Council 模式建议填写 `--perspective`:
    
    ```bash
    python3 skills/agent-team-manager/scripts/agent_team_manager.py member-add "投资评审会" \
      --id risk \
      --role "风险分析师" \
      --perspective "风险与合规"
    ```
    
    ### 修改成员
    
    ```bash
    python3 skills/agent-team-manager/scripts/agent_team_manager.py member-update "文档深度分析" analyzer \
      --task "重点分析论证链、信息缺口和潜在偏见" \
      --system-prompt "你是批判性分析专家,输出问题、证据和风险。"
    ```
    
    ### 修改现有角色提示词(强约束)
    
    凡是用户表达以下意图,统一按“修改现有角色提示词”处理:
    
    - 优化角色提示词
    - 改 prompt / 改系统提示词
    - 调整角色设定、语气、边界、输出要求
    - 让某个成员“更专业 / 更严格 / 更像某类专家”
    
    强制执行顺序如下。
    
    #### 1. 先读取当前真实配置,不允许凭印象修改
    
    ```bash
    python3 skills/agent-team-manager/scripts/agent_team_manager.py info "文档深度分析"
    ```
    
    必须先看目标成员当前是否存在 `task`、`system_prompt`,再决定修改哪个字段。
    禁止在未读取当前配置前直接生成 `member-update` 命令。
    
    #### 2. 字段归属判定规则
    
    1. 若目标成员只有 `task`,没有 `system_prompt`
    2. 说明该团队把提示词主体直接存放在 `task`
    3. 此时用户说“改提示词”,默认优先修改 `--task`
    4. 若目标成员已有 `system_prompt`
    5. 涉及角色人格、专家身份、口吻、原则、边界、长期行为约束时,优先修改 `--system-prompt`
    6. 涉及具体工作内容、执行步骤、输出结构、交付格式、检查项时,优先修改 `--task`
    7. 若当前 `task` 本身是一整段提示词式文本,且用户想做系统化重构,应同时修改 `--task` 与 `--system-prompt`
    
    #### 3. 执行规范
    
    - 修改必须落库,最终动作一定是执行 `member-update`
    - `member-update` 的正确形式是:
      `python3 skills/agent-team-manager/scripts/agent_team_manager.py member-update "团队名" 成员ID [flags]`
    - 成员标识使用位置参数 `member_ref`,不要写成 `--id`
    - 修改完成后,建议再次执行 `info "团队名"` 复核结果
    - 如果用户要求“修改提示词”,但当前配置里只有 `task`,不要只改 `system_prompt`
    - 如果用户要求“系统提示词更专业”,但旧的长提示还残留在 `task` 中,应判断是否需要同步精简 `task`
    
    示例:当前成员只有 `task`,没有 `system_prompt`
    
    ```bash
    python3 skills/agent-team-manager/scripts/agent_team_manager.py member-update "小红书文案团队" reviewer \
      --task "审核小红书文案,检查敏感词、风格统一、内容完整、平台规范和可读性;输出审核结论、修改建议、最终发布版与发布提醒。"
    ```
    
    示例:明确修改系统提示词
    
    ```bash
    python3 skills/agent-team-manager/scripts/agent_team_manager.py member-update "小红书文案团队" reviewer \
      --system-prompt "你是严格的小红书内容审核专家,优先识别违规风险和夸大表达,输出结论必须清晰、克制、可执行。"
    ```
    
    示例:把旧的提示词式 task 拆成“任务 + 系统提示词”
    
    ```bash
    python3 skills/agent-team-manager/scripts/agent_team_manager.py member-update "小红书文案团队" reviewer \
      --task "审核小红书文案并输出审核结论、修改建议、最终发布版与发布提醒。" \
      --system-prompt "你是严格的小红书内容审核专家,重点检查敏感词、极限词、医疗宣称、风格统一、内容完整和平台规范。"
    ```
    
    #### 4. 专业处理原则
    
    - 目标是“修改有效配置”,不是“输出一段看起来更好的文案”
    - 先识别现有数据结构,再决定改哪个字段
    - 以最小必要修改为原则,避免只新增字段而保留旧冲突内容
    - 若用户未指定字段名,“提示词”一词要结合当前存储结构解释,不得机械等同于 `system_prompt`
    
    ### 删除成员
    
    ```bash
    python3 skills/agent-team-manager/scripts/agent_team_manager.py member-delete "文档深度分析" summarizer
    ```
    
    ## 团队模型配置
    
    ### 查看当前配置
    
    ```bash
    python3 skills/agent-team-manager/scripts/agent_team_manager.py config "文档深度分析"
    ```
    
    ### 设置自定义模型
    
    ```bash
    python3 skills/agent-team-manager/scripts/agent_team_manager.py config-set "文档深度分析" \
      --provider zhipu \
      --model glm-5 \
      --temperature 0.5 \
      --max-tokens 8192
    ```
    
    ### 重置为全局默认
    
    ```bash
    python3 skills/agent-team-manager/scripts/agent_team_manager.py config-reset "文档深度分析"
    ```
    
    ## 参数说明
    
    - `mode` 仅支持 `pipeline`、`graph`、`council`
    - `--enable-skills` 开启后,子 Agent 可读取并使用 `skills/*`
    - `--cross-review` / `--no-cross-review` 仅对 `council` 模式有意义
    - `depends_on` 仅对 `graph` 模式有意义
    - `perspective` 主要用于 `council` 模式
    - `task` 会作为执行阶段的 `# Your Task` 传给子 Agent,是实际任务说明
    - `system_prompt` 是角色长期人格/职责设定
    - 如果 `system_prompt` 为空,系统会基于 `role + task` 自动生成默认系统提示词
    - 如果 `system_prompt` 有值,会直接作为系统消息使用;但 `task` 仍然会继续传入执行提示中
    - 因此:已有成员只有 `task` 时,优先更新 `task`;需要稳定角色口吻/边界时,再补或修改 `system_prompt`
    - 很多历史团队把整段“提示词式描述”直接写进了 `task`,这不是脚本失效,而是数据本来就这样存的
    - 所以“提示词改了没生效”时,优先检查是不是旧提示还躺在 `task` 里
    
    ## 内置模板
    
    当前内置:
    
    - `document-analysis`:文档深度分析,`pipeline` 模式,默认开启技能系统
    
    ## 注意事项
    
    - 团队名称必须唯一
    - 成员 ID 在同一团队内必须唯一
    - 修改成员本质上会读取团队详情后整体回写 `agents`
    - 如果团队启用了专属模型,执行 `workflow_run(team_name="团队名", goal="...")` 时会自动继承该模型配置
    - 如果 CountBot 后端未启动,脚本会直接报连接失败
    
  • README.mdupstream ↗

    Contents are not inlined in the catalog. Open the upstream file.

Plain-text copy, no login: /skill-bundles/countbot-ai-countbot-workspace-skills-agent-team-manager.md

// REPO STATS

785 stars
14 open issues
Last commit: 16d ago