agi-farm

Builds a complete multi-agent AI team on OpenClaw. One wizard, full team.

Commands

Command	What it does
INLINECODE0	Full wizard — agents, workspace, crons, bundle, GitHub
INLINECODE1

`/agi-farm setup`

Ask one question at a time. Do not proceed until confirmed.

Step 1 — Team name

"What should we call your team? (e.g. NovaCorp, TradingDesk — default: MyTeam)"

Store as TEAM_NAME.

Step 2 — Orchestrator name

"What's your orchestrator's name? (default: Cooper)"

Store as ORCHESTRATOR_NAME.

Step 3 — Team size

"How many agents?
3 — Minimal: Orchestrator + Researcher + Builder
5 — Standard: adds QA + Content
11 — Full stack: complete AGI system (recommended)"

Store as PRESET.

Step 3.5 — Domain

"What domain? software / trading / research / general (default) / custom"

If custom: ask for one-phrase description. Store as DOMAIN.

Step 3.6 — Custom agents (PRESET 3 or 5 only)

"Add a custom agent? (yes/no, default: no)"

If yes, collect per agent: id, name, emoji, role, goal. Max 3 custom agents.
Append to roster in Step 7 with "template": "generic".

Step 4 — Frameworks

"Collaboration frameworks? autogen / crewai / langgraph / all / none"

Store as FRAMEWORKS list. all → ["autogen", "crewai", "langgraph"].

Step 5 — GitHub

"Create a GitHub repo for the bundle? yes / no"

Store as CREATE_GITHUB.

Step 6 — Confirm

Show summary, ask "Shall I proceed? (yes/no)". If no → restart Step 1.

Step 7 — Write `team.json`

CODEBLOCK0

Use the openclaw agents list output to assign each agent a model appropriate for
its role. Write resolved model strings directly into the "model" fields.

Model selection cheat sheet (based on openclaw agents list --json output):

Role	Recommended tier	Why
Orchestrator	High-capability (e.g. `sonnet`, `opus`)	Needs broad reasoning, delegation judgment
Solution Architect / Researcher

Tip: assign opus or sonnet to roles that make decisions; use flash/glm-4.7-flash for high-frequency reviewers to manage cost.

3-agent roster:
CODEBLOCK1

5-agent: add to 3-agent roster:
CODEBLOCK2

11-agent roster:

[
  {"id":"main",  "name":"<ORCHESTRATOR_NAME>","emoji":"🦅","role":"Orchestrator",            "goal":"Orchestrate specialists, delegate tasks, synthesize results",                          "model":"<MODEL>","workspace":"."},
  {"id":"sage",  "name":"Sage",              "emoji":"🔮","role":"Solution Architect",       "goal":"Design robust, scalable architectures",                                               "model":"<MODEL>","workspace":"solution-architect"},
  {"id":"forge", "name":"Forge",             "emoji":"⚒️","role":"Implementation Engineer", "goal":"Implement clean, well-tested code efficiently",                                        "model":"<MODEL>","workspace":"implementation-engineer"},
  {"id":"pixel", "name":"Pixel",             "emoji":"🐛","role":"Debugger",                "goal":"Find the true root cause of any bug or failure",                                       "model":"<MODEL>","workspace":"debugger"},
  {"id":"vista", "name":"Vista",             "emoji":"🔭","role":"Business Analyst",        "goal":"Research deeply and surface the insights that matter most",                            "model":"<MODEL>","workspace":"business-analyst"},
  {"id":"cipher","name":"Cipher",            "emoji":"🔊","role":"Knowledge Curator",       "goal":"Curate and surface knowledge so the team never forgets",                              "model":"<MODEL>","workspace":"knowledge-curator"},
  {"id":"vigil", "name":"Vigil",             "emoji":"🛡️","role":"QA Engineer",             "goal":"Ensure every output meets quality standards",                                         "model":"<MODEL>","workspace":"quality-assurance"},
  {"id":"anchor","name":"Anchor",            "emoji":"⚓", "role":"Content Specialist",     "goal":"Craft clear content that communicates complex ideas simply",                           "model":"<MODEL>","workspace":"content-specialist"},
  {"id":"lens",  "name":"Lens",              "emoji":"📡","role":"Multimodal Specialist",   "goal":"Extract meaning from images, documents, and multimodal inputs",                       "model":"<MODEL>","workspace":"multimodal-specialist"},
  {"id":"evolve","name":"Evolve",            "emoji":"🔄","role":"Process Improvement Lead","goal":"Make the team better systematically through continuous improvement",                   "model":"<MODEL>","workspace":"process-improvement"},
  {"id":"nova",  "name":"Nova",              "emoji":"🧪","role":"R&D Lead",                "goal":"Turn hypotheses into proven capabilities through structured experimentation",          "model":"<MODEL>","workspace":"r-and-d"}
]

Step 8 — Generate workspace files

CODEBLOCK4

Step 9 — Create OpenClaw agents

For each agent except main (skip if already exists):

CODEBLOCK5

Use agent["model"] from team.json directly.

Step 10 — Register cron jobs

CODEBLOCK6

Timezone is read automatically from OpenClaw config. Skips any cron that already exists.

Step 11 — Install frameworks

For each framework in FRAMEWORKS:

CODEBLOCK7

Step 12 — GitHub (if chosen)

CODEBLOCK8

Step 13 — Commit workspace

CODEBLOCK9

Step 14 — Initialize registries + health check

CODEBLOCK10

Step 15 — Done

CODEBLOCK11

`/agi-farm status`

CODEBLOCK12

`/agi-farm rebuild`

CODEBLOCK13

INLINECODE42 skips files that already exist, preserving manual edits.
Add --force (remove --no-overwrite) to overwrite everything.

`/agi-farm export`

CODEBLOCK14

`/agi-farm dashboard`

React + SSE ops room. File-watcher pushes live data to the browser in ~350ms on any workspace .json or .md change. Runs as a persistent macOS LaunchAgent — always on, auto-restarts on crash.

Architecture

CODEBLOCK15

Data sources (all real-time from workspace files)

Field	Source file	Refresh
tasks, taskcounts, slaat_risk	INLINECODE49	instant
agents (inbox, perf, status)

LaunchAgent (always-on)

The dashboard is registered as ai.coopercorp.dashboard and starts automatically at login.

CODEBLOCK16

URL: http://localhost:8080

Rebuild React frontend

CODEBLOCK17

Full reference: references/dashboard.md

`/agi-farm dispatch`

CODEBLOCK18

Fires agent sessions for pending tasks, handles HITL notifications, stale task
resets, rate-limit backoff, and dependency checking. Cron (every 1 min):

* * * * * python3 ~/.openclaw/skills/agi-farm/scripts/auto-dispatch.py --execute \
            >> ~/.openclaw/workspace/logs/auto-dispatch.log 2>&1

Troubleshooting

Setup issues

Symptom	Fix
INLINECODE64 fails with INLINECODE65	Run INLINECODE66
INLINECODE67 says agent already exists

Safe to ignore — skip that agent | | gh repo create fails | Run gh auth login first | | Cron registration shows 0 crons added | Run openclaw cron list to check for duplicates; use --force flag on re-register | | git commit fails in Step 13 | Run git config --global user.email and set name/email first |

Runtime issues

Symptom	Fix
Auto-dispatcher fires but agents don't respond	Check `logs/auto-dispatch.log`; verify `openclaw agents list` shows agents
Dashboard shows stale data

Restart LaunchAgent: launchctl stop ai.coopercorp.dashboard && launchctl start ai.coopercorp.dashboard | | TASKS.json parse error | Validate JSON: python3 -m json.tool ~/.openclaw/workspace/TASKS.json | | Agent stuck >30 min | Check broadcast.md for [BLOCKED] tags; reassign task manually | | Rate-limit backoff too aggressive | Edit RATE_LIMIT_BACKOFF_MIN in scripts/auto-dispatch.py (default: 10 min) | | openclaw not found in cron | Set OPENCLAW_BIN=/path/to/openclaw in the cron environment, or add PATH=/opt/homebrew/bin:$PATH |

Recovery

CODEBLOCK20

agi-farm

在 OpenClaw 上构建一个完整的多智能体 AI 团队。一个向导，完整的团队。

命令

命令	功能
/agi-farm setup	完整向导 — 智能体、工作区、定时任务、捆绑包、GitHub
/agi-farm status

/agi-farm setup

每次只问一个问题。在确认之前不要继续。

步骤 1 — 团队名称

你的团队叫什么名字？（例如 NovaCorp、TradingDesk — 默认：MyTeam）

存储为 TEAM_NAME。

步骤 2 — 协调器名称

你的协调器叫什么名字？（默认：Cooper）

存储为 ORCHESTRATOR_NAME。

步骤 3 — 团队规模

需要多少个智能体？
3 — 最小化：协调器 + 研究员 + 构建者
5 — 标准：增加 QA + 内容
11 — 全栈：完整的 AGI 系统（推荐）

存储为 PRESET。

步骤 3.5 — 领域

什么领域？软件 / 交易 / 研究 / 通用（默认）/ 自定义

如果选择自定义：要求提供一句话描述。存储为 DOMAIN。

步骤 3.6 — 自定义智能体（仅限 PRESET 3 或 5）

添加自定义智能体？（是/否，默认：否）

如果是，为每个智能体收集：id、name、emoji、role、goal。最多 3 个自定义智能体。
在步骤 7 中追加到名册，使用 template: generic。

步骤 4 — 框架

协作框架？autogen / crewai / langgraph / all / none

存储为 FRAMEWORKS 列表。all → [autogen, crewai, langgraph]。

步骤 5 — GitHub

为捆绑包创建 GitHub 仓库？是 / 否

存储为 CREATE_GITHUB。

步骤 6 — 确认

显示摘要，询问是否继续？（是/否）。如果否 → 重新开始步骤 1。

步骤 7 — 写入 team.json

bash
mkdir -p ~/.openclaw/workspace/agi-farm-bundle/
openclaw agents list --json # 使用输出为每个角色分配适当的模型

使用 openclaw agents list 输出为每个智能体分配适合其角色的模型。将解析后的模型字符串直接写入 model 字段。

模型选择速查表（基于 openclaw agents list --json 输出）：

角色	推荐层级	原因
协调器	高能力（例如 sonnet、opus）	需要广泛的推理、委派判断
解决方案架构师 / 研究员

提示：将 opus 或 sonnet 分配给做决策的角色；使用 flash/glm-4.7-flash 给高频审查者以控制成本。

3 智能体名册：
json
{teamname:NAME>,orchestratorname:NAME>,preset:3,
domain:,frameworks:JSON>,createdat:,
agents:[
{id:main, name:,emoji:🦅,role:Orchestrator, goal:协调团队、委派任务、综合结果, model:,workspace:.},
{id:researcher, name:Sage, emoji:🔮,role:Researcher, goal:深入研究并揭示最重要的洞察, model:,workspace:researcher},
{id:builder, name:Forge, emoji:⚒️,role:Builder, goal:干净高效地实现解决方案, model:,workspace:builder}
]}

5 智能体： 在 3 智能体名册基础上增加：
json
{id:qa, name:Vigil, emoji:🛡️,role:QA Engineer, goal:确保每个输出都符合质量标准,model:,workspace:qa},
{id:content,name:Anchor,emoji:⚓, role:Content Specialist,goal:制作清晰的内容，用简单的方式传达复杂想法,model:,workspace:content}

11 智能体名册：
json
[
{id:main, name:,emoji:🦅,role:Orchestrator, goal:协调专家、委派任务、综合结果, model:,workspace:.},
{id:sage, name:Sage, emoji:🔮,role:Solution Architect, goal:设计稳健、可扩展的架构, model:,workspace:solution-architect},
{id:forge, name:Forge, emoji:⚒️,role:Implementation Engineer, goal:高效实现干净、经过良好测试的代码, model:,workspace:implementation-engineer},
{id:pixel, name:Pixel, emoji:🐛,role:Debugger, goal:找到任何错误或失败的真实根本原因, model:,workspace:debugger},
{id:vista, name:Vista, emoji:🔭,role:Business Analyst, goal:深入研究并揭示最重要的洞察, model:,workspace:business-analyst},
{id:cipher,name:Cipher, emoji:🔊,role:Knowledge Curator, goal:策划和呈现知识，让团队永不遗忘, model:,workspace:knowledge-curator},
{id:vigil, name:Vigil, emoji:🛡️,role:QA Engineer, goal:确保每个输出都符合质量标准, model:,workspace:quality-assurance},
{id:anchor,name:Anchor, emoji:⚓, role:Content Specialist, goal:制作清晰的内容，用简单的方式传达复杂想法, model:,workspace:content-specialist},
{id:lens, name:Lens, emoji:📡,role:Multimodal Specialist, goal:从图像、文档和多模态输入中提取意义, model:,workspace:multimodal-specialist},
{id:evolve,name:Evolve, emoji:🔄,role:Process Improvement Lead,goal:通过持续改进系统性地让团队变得更好, model:,workspace:process-improvement},
{id:nova, name:Nova, emoji:🧪,role:R&D Lead, goal:通过结构化实验将假设转化为

敏捷农场

敏捷农场

agi-farm

Commands

/agi-farm setup

Step 1 — Team name

Step 2 — Orchestrator name

Step 3 — Team size

Step 3.5 — Domain

Step 3.6 — Custom agents (PRESET 3 or 5 only)

Step 4 — Frameworks

Step 5 — GitHub

Step 6 — Confirm

Step 7 — Write team.json

Step 8 — Generate workspace files

Step 9 — Create OpenClaw agents

Step 10 — Register cron jobs

Step 11 — Install frameworks

Step 12 — GitHub (if chosen)

Step 13 — Commit workspace

Step 14 — Initialize registries + health check

Step 15 — Done

/agi-farm status

/agi-farm rebuild

/agi-farm export

/agi-farm dashboard

Architecture

Data sources (all real-time from workspace files)

LaunchAgent (always-on)

Rebuild React frontend

/agi-farm dispatch

Troubleshooting

Setup issues

Runtime issues

Recovery

agi-farm

命令

/agi-farm setup

步骤 1 — 团队名称

步骤 2 — 协调器名称

步骤 3 — 团队规模

步骤 3.5 — 领域

步骤 3.6 — 自定义智能体（仅限 PRESET 3 或 5）

步骤 4 — 框架

步骤 5 — GitHub

步骤 6 — 确认

步骤 7 — 写入 team.json

标签

通过对话安装

方式一：安装 SkillHub 和技能

方式二：设置 SkillHub 为优先技能安装源

通过命令行安装

下载

相关推荐

self-improvement

self-improvement

self-improvement

self-improvement

`/agi-farm setup`

Step 7 — Write `team.json`

`/agi-farm status`

`/agi-farm rebuild`

`/agi-farm export`

`/agi-farm dashboard`

`/agi-farm dispatch`