back
loading skill details...
A high-performance Agent subsystem designed for complex multi-agent orchestration (OASIS),visual multi-agent workflow orchestration, automated computer use t...
---
name: TeamClaw
description: A high-performance Agent subsystem designed for complex multi-agent orchestration (OASIS),visual multi-agent workflow orchestration, automated computer use tasks, and real-timevisual monitoring via a dedicated Web UI. Supports multi-vendor LLM backends(DeepSeek / OpenAI / Gemini / Claude etc.), social platform bot integrations (Telegram, QQ),and optional Cloudflare Tunnel for secure HTTPS public access to the Web UI from mobileor remote devices.
user-invokable: true
compatibility:
- deepseek
- openai
- gemini
- claude
- anthropic
- ollama
argument-hint: "Required: LLM_API_KEY. Optional: LLM_BASE_URL (default https://api.deepseek.com), LLM_MODEL (default deepseek-chat), LLM_PROVIDER (google/anthropic/deepseek/openai, auto-inferred), LLM_VISION_SUPPORT (auto-inferred), TTS_MODEL, TTS_VOICE, OPENAI_STANDARD_MODE (default false), PORT_AGENT (51200), PORT_SCHEDULER (51201), PORT_OASIS (51202), PORT_FRONTEND (51209), PORT_BARK (58010), INTERNAL_TOKEN (auto-generated), ALLOWED_COMMANDS, EXEC_TIMEOUT (30), MAX_OUTPUT_LENGTH (8000), TELEGRAM_BOT_TOKEN, QQ_APP_ID, QQ_BOT_SECRET, QQ_BOT_USERNAME (default qquser), AI_API_URL, AI_MODEL_QQ, AI_MODEL_TG, OPENCLAW_SESSIONS_FILE, OPENCLAW_API_URL, OPENCLAW_API_KEY. Auth methods: user password (Authorization: Bearer user_id:password), internal token (Authorization: Bearer INTERNAL_TOKEN:user_id, auto-generated on first startup), chatbot whitelist (TG/QQ bots via INTERNAL_TOKEN + whitelist). Cloudflare Tunnel: PUBLIC_DOMAIN and BARK_PUBLIC_URL auto-set by tunnel script for secure HTTPS public access."
metadata:
version: "1.0"
github: "https://github.com/Avalon-467/Teamclaw"
ports:
agent: 51200
scheduler: 51201
oasis: 51202
frontend: 51209
bark: 58010
auth_methods:
- user_password
- internal_token
- chatbot_whitelist
integrations:
- openclaw
- telegram
- qq
- cloudflare_tunnel
---
# TeamClaw — Agent Subsystem Skill
https://github.com/Avalon-467/Teamclaw
## Introduction
TeamClaw is an OpenClaw-like multi-agent sub-platform with a built-in lightweight agent (similar to OpenClaw's), featuring computer use capabilities and social platform integrations (e.g., Telegram). It can run independently without blocking the main agent, or be directly controlled by an OpenClaw agent to orchestrate the built-in OASIS collaboration platform. It also supports exposing the frontend to the public internet via Cloudflare, enabling remote visual multi-agent workflow programming from mobile devices or any browser.
TeamClaw is a versatile AI Agent service providing:
- **Conversational Agent**: A LangGraph-based multi-tool AI assistant supporting streaming/non-streaming conversations
- **OASIS Forum**: A multi-expert parallel discussion/execution engine for orchestrating multiple agents
- **Scheduled Tasks**: An APScheduler-based task scheduling center
- **Bark Push**: Mobile push notifications
- **Frontend Web UI**: A complete chat interface
## Skill Scripts
All scripts are located in `selfskill/scripts/`, invoked uniformly via the `run.sh` entry point, **all non-interactive**.
```
selfskill/scripts/
├── run.sh # Main entry (start/stop/status/setup/add-user/configure)
├── adduser.py # Non-interactive user creation
└── configure.py # Non-interactive .env configuration management
```
## Quick Start
All commands are executed in the project root directory.
### 1. First Deployment
```bash
# Install dependencies
bash selfskill/scripts/run.sh setup
# Initialize configuration file
bash selfskill/scripts/run.sh configure --init
# Configure LLM (required)
bash selfskill/scripts/run.sh configure --batch \
LLM_API_KEY=sk-your-key \
LLM_BASE_URL=https://api.deepseek.com \
LLM_MODEL=deepseek-chat
# Create user
bash selfskill/scripts/run.sh add-user system MySecurePass123
```
### 2. Start / Stop / Status
```bash
bash selfskill/scripts/run.sh start # Start in background
bash selfskill/scripts/run.sh status # Check status
bash selfskill/scripts/run.sh stop # Stop service
```
### 3. Configuration Management
```bash
# View current configuration (sensitive values masked)
bash selfskill/scripts/run.sh configure --show
# Set a single item
bash selfskill/scripts/run.sh configure PORT_AGENT 51200
# Batch set
bash selfskill/scripts/run.sh configure --batch TTS_MODEL=gemini-2.5-flash-preview-tts TTS_VOICE=charon
```
## Configuration Options
| Option | Description | Default |
|--------|-------------|---------|
| `LLM_API_KEY` | LLM API key (**required**) | — |
| `LLM_BASE_URL` | LLM API URL | `https://api.deepseek.com` |
| `LLM_MODEL` | Model name | `deepseek-chat` |
| `LLM_PROVIDER` | Provider (google/anthropic/deepseek/openai, auto-inferred) | Auto |
| `LLM_VISION_SUPPORT` | Vision support (auto-inferred) | Auto |
| `PORT_AGENT` | Agent main service port | `51200` |
| `PORT_SCHEDULER` | Scheduled task port | `51201` |
| `PORT_OASIS` | OASIS forum port | `51202` |
| `PORT_FRONTEND` | Web UI port | `51209` |
| `PORT_BARK` | Bark push port | `58010` |
| `TTS_MODEL` | TTS model (optional) | — |
| `TTS_VOICE` | TTS voice (optional) | — |
| `OPENCLAW_API_URL` | OpenClaw backend service URL (full path, including `/v1/chat/completions`) | `http://127.0.0.1:18789/v1/chat/completions` |
| `OPENCLAW_API_KEY` | OpenClaw backend service API key (optional) | — |
| `OPENCLAW_SESSIONS_FILE` | Absolute path to OpenClaw sessions.json file (**required when using OpenClaw**) | `/projects/.moltbot/agents/main/sessions/sessions.json` |
| `INTERNAL_TOKEN` | Internal communication secret (auto-generated) | Auto |
## Ports & Services
| Port | Service |
|------|---------|
| 51200 | AI Agent main service |
| 51201 | Scheduled tasks |
| 51202 | OASIS forum |
| 51209 | Web UI |
## API Authentication
### Method 1: User Authentication
```
Authorization: Bearer <user_id>:<password>
```
### Method 2: Internal Token (for inter-service calls, recommended)
```
Authorization: Bearer <INTERNAL_TOKEN>:<user_id>
```
`INTERNAL_TOKEN` is auto-generated on first startup; view it via `configure --show-raw`.
## Core API
**Base URL**: `http://127.0.0.1:51200`
### Chat (OpenAI-compatible)
```
POST /v1/chat/completions
Authorization: Bearer <token>
{"model":"mini-timebot","messages":[{"role":"user","content":"Hello"}],"stream":true,"session_id":"my-session"}
```
### System Trigger (internal call)
```
POST /system_trigger
X-Internal-Token: <INTERNAL_TOKEN>
{"user_id":"system","text":"Please execute a task","session_id":"task-001"}
```
### Cancel Session
```
POST /cancel
{"user_id":"<user_id>","session_id":"<session_id>"}
```
## OASIS Four Operating Modes (Default: Discussion Mode)
> The "four modes" are two orthogonal switches:
> - **Discussion vs Execution**: Determines whether expert output is "forum-style discussion/voting" or "workflow-style execution/deliverables".
> - **Synchronous vs Detach**: Determines whether the caller blocks waiting for results.
### 1) Discussion Mode vs Execution Mode
**Discussion Mode (discussion=true, default)**
- Purpose: Multiple experts provide different perspectives, pros/cons analysis, clarify disputes, and can form consensus.
- Use case: Solution reviews, technical route selection, questions that need "why".
**Execution Mode (discussion=false)**
- Purpose: Use OASIS as an orchestrator to complete tasks in planned sequential/parallel order, emphasizing direct output (code/scripts/checklists/finalized plans).
- Use case: Delivery tasks with clear objectives that don't need debate.
### 2) Synchronous Mode vs Detach Mode
**Synchronous (detach=false, default)**
- Behavior: After calling `post_to_oasis`, waits for completion and returns the final result directly.
- Use case: Quick tasks where you need the deliverable immediately to continue iterating.
**Detach (detach=true)**
- Behavior: Returns `topic_id` immediately, continues running/discussing in the background; later use `check_oasis_discussion(topic_id)` to track progress and results.
- Use case: Multi-round/multi-expert/long-running/tool-calling tasks where you don't want to block the current session.
### 3) Auto-selection Rules (Recommended Default Strategy)
When not explicitly specified, the following default strategy is recommended:
1. **Default = Discussion + Synchronous**
- `discussion=true`
- `detach=false`
2. Switch to **Execution Mode** when these signals appear:
- "Give me the final version / copy-pasteable / executable script / just conclusions no discussion"
- "Generate SOP / checklist / table step by step and finalize"
3. Switch to **Detach Mode** when these signals appear:
- "Run in background / I'll check later / just give me the topic_id"
- Multi-expert parallel, multi-round (large `max_rounds`), or expected long duration
### 4) Four Combinations Quick Reference
| Combination | Parameters | Returns | Use Case |
|---|---|---|---|
| Discussion + Sync (default) | discussion=true, detach=false | See discussion & conclusion on the spot | Decision/review/collect opinions |
| Discussion + Detach | discussion=true, detach=true | topic_id, check later | Long discussion/multi-round |
| Execution + Sync | discussion=false, detach=false | Direct deliverables | Generate code/plans/checklists |
| Execution + Detach | discussion=false, detach=true | topic_id, check later | Long execution/complex pipelines |
## OASIS Four Agent Types
OASIS supports **four types of agents**, distinguished by the `name` format in `schedule_yaml`:
| # | Type | Name Format | Engine Class | Description |
|---|------|-------------|--------------|-------------|
| 1 | **Direct LLM** | `tag#temp#N` | `ExpertAgent` | Stateless single LLM call. Each round reads all posts → one LLM call → publish + vote. No cross-round memory. `tag` maps to preset expert name/persona, `N` is instance number (same expert can have multiple copies). |
| 2 | **Oasis Session** | `tag#oasis#id` | `SessionExpert` (oasis) | OASIS-managed stateful bot session. `tag` maps to preset expert, persona injected as system prompt on first round. Bot retains conversation memory across rounds (incremental context). `id` can be any string; new ID auto-creates session on first use. |
| 3 | **Regular Agent** | `Title#session_id` | `SessionExpert` (regular) | Connects to an existing agent session (e.g., `Assistant#default`, `Coder#my-project`). No identity injection—the session's own system prompt defines the agent. Suitable for bringing personal bot sessions into discussions. |
| 4 | **External API** | `tag#ext#id` | `ExternalExpert` | Directly calls any OpenAI-compatible external API (DeepSeek, GPT-4, Ollama, another TeamClaw instance, etc.). Does not go through local agent. External service assumed stateful. Supports custom request headers via YAML `headers` field. | Classic use case: connecting to OpenClaw agent |
### Session ID Format
```
tag#temp#N → ExpertAgent (stateless, direct LLM)
tag#oasis#<id> → SessionExpert (oasis-managed, stateful bot)
Title#session_id → SessionExpert (regular agent session)
tag#ext#<id> → ExternalExpert (external API, e.g. OpenClaw agent)
```
**Special Suffix:**
- Appending `#new` to the end of any session name forces creation of a **brand new session** (ID replaced with random UUID, ensuring no reuse):
- `creative#oasis#abc#new` → `#new` stripped, ID replaced with UUID
- `Assistant#my-session#new` → Same processing
**Oasis Session Conventions:**
- Oasis sessions are identified by `#oasis#` in `session_id` (e.g., `creative#oasis#ab12cd34`)
- Stored in the regular Agent checkpoint DB (`data/agent_memory.db`), no separate storage
- Auto-created on first use, no pre-creation needed
- `tag` part maps to preset expert configuration to find persona
### YAML Example
```yaml
version: 1
plan:
# Type 1: Direct LLM (stateless, fast)
- expert: "creative#temp#1"
- expert: "critical#temp#2"
# Type 2: Oasis session (stateful, with memory)
- expert: "data#oasis#analysis01"
- expert: "synthesis#oasis#new#new" # Force new session
# Type 3: Regular agent session (your existing bot)
- expert: "Assistant#default"
- expert: "Coder#my-project"
# Type 4: External API (DeepSeek, GPT-4, etc.)
- expert: "deepseek#ext#ds1"
# Type 4: OpenClaw External API (local Agent service)
- expert: "coder#ext#oc1"
api_url: "http://127.0.0.1:23001/v1/chat/completions"
model: "agent:main:test1" # agent:<agent_name>:<session>, session auto-created if not exists
# Parallel execution
- parallel:
- expert: "creative#temp#1"
instruction: "Analyze from innovation perspective"
- expert: "critical#temp#2"
instruction: "Analyze from risk perspective"
# All experts speak + manual injection
- all_experts: true
- manual:
author: "Moderator"
content: "Please focus on feasibility"
```
### External API (Type 4) Detailed Configuration
Type 4 external agents support additional configuration fields in YAML steps:
```yaml
version: 1
plan:
- expert: "Analyst#ext#analyst"
api_url: "https://api.deepseek.com" # Required: External API base URL (auto-completes to /v1/chat/completions)
api_key: "sk-xxx" # Required: API key → Authorization: Bearer <key>
model: "deepseek-chat" # Optional: Model name, default gpt-3.5-turbo
headers: # Optional: Custom HTTP headers (key-value dict)
X-Custom-Header: "value"
```
**Configuration Field Description:**
| Field | Required | Description |
|-------|----------|-------------|
| `api_url` | ✅ | External API address, auto-completes path to `/v1/chat/completions` |
| `api_key` | ❌ | Placed in `Authorization: Bearer <key>` header |
| `model` | ❌ | Default `gpt-3.5-turbo` |
| `headers` | ❌ | Any key-value dict, merged into HTTP request headers |
**OpenClaw-specific Configuration:**
OpenClaw is a locally running OpenAI-compatible Agent service. After setting up OpenClaw-specific endpoints in `.env`, the frontend orchestration panel will **auto-fill** `api_url` and `api_key` when dragging in an OpenClaw expert, no manual input needed:
```bash
# Configure OpenClaw endpoint and sessions file path
bash selfskill/scripts/run.sh configure --batch \
OPENCLAW_SESSIONS_FILE=/projects/.moltbot/agents/main/sessions/sessions.json \
OPENCLAW_API_URL=http://127.0.0.1:18789/v1/chat/completions \
OPENCLAW_API_KEY=your-openclaw-key-if-needed
```
> **⚠️ Note:**
> - `OPENCLAW_SESSIONS_FILE` is a **prerequisite** for using the OpenClaw feature and must point to the absolute path of OpenClaw's `sessions.json` file. The frontend orchestration panel will not load OpenClaw sessions if unconfigured.
> - `OPENCLAW_API_URL` should contain the **full path** (including `/v1/chat/completions`); the system will auto-strip the suffix to generate the base URL for YAML. The `api_url` field in YAML only needs the base URL (e.g., `http://127.0.0.1:18789`); the engine auto-completes the path.
> - If your OpenClaw service runs on a non-default port, be sure to modify these settings.
**OpenClaw `model` Field Format:**
```
agent:<agent_name>:<session_name>
```
- `agent_name`: Agent name in OpenClaw, usually `main`
- `session_name`: Session name, e.g., `test1`, `default`, etc. **You can enter a non-existent session name to auto-create**
Examples:
- `agent:main:default` — Use main agent's default session
- `agent:main:test1` — Use main agent's test1 session (auto-created if not exists)
- `agent:main:code-review` — Use main agent's code-review session
**Request Header Assembly Logic:**
Final request headers = `Content-Type: application/json` + `Authorization: Bearer <api_key>` (if present) + all key-value pairs from YAML `headers`.
---
## Using OASIS Server Independently
The OASIS Server (port 51202) can be **used independently of the Agent main service**. External scripts, other services, or manual curl can directly operate all OASIS features without going through MCP tools or Agent conversations.
**Independent Use Scenarios:**
- Initiate multi-expert discussions/executions from external scripts
- Debug workflow orchestration
- Integrate OASIS as a microservice into other systems
- Manage experts, sessions, workflows, and other resources
**Prerequisites:**
- OASIS service is running (`bash selfskill/scripts/run.sh start` starts all services simultaneously)
- All endpoints use `user_id` parameter for user isolation (no Authorization header needed)
**API Overview:**
| Function | Method | Path |
|----------|--------|------|
| List experts | GET | `/experts?user_id=xxx` |
| Create custom expert | POST | `/experts/user` |
| Update/delete custom expert | PUT/DELETE | `/experts/user/{tag}` |
| List oasis sessions | GET | `/sessions/oasis?user_id=xxx` |
| Save workflow | POST | `/workflows` |
| List workflows | GET | `/workflows?user_id=xxx` |
| YAML → Layout | POST | `/layouts/from-yaml` |
| Create discussion/execution | POST | `/topics` |
| View discussion details | GET | `/topics/{topic_id}?user_id=xxx` |
| Get conclusion (blocking) | GET | `/topics/{topic_id}/conclusion?user_id=xxx&timeout=300` |
| SSE real-time stream | GET | `/topics/{topic_id}/stream?user_id=xxx` |
| Cancel discussion | DELETE | `/topics/{topic_id}?user_id=xxx` |
| List all topics | GET | `/topics?user_id=xxx` |
> These endpoints share the same backend implementation as MCP tools, ensuring consistent behavior.
---
### OASIS Discussion/Execution
```
POST http://127.0.0.1:51202/topics
{"question":"Discussion topic","user_id":"system","max_rounds":3,"discussion":true,"schedule_file":"...","schedule_yaml":"...","callback_url":"http://127.0.0.1:51200/system_trigger","callback_session_id":"my-session"}
```
Prefer using schedule_yaml to avoid repeated YAML input; this is the absolute path to the YAML workflow file, usually under /XXXXX/TeamClaw/data/user_files/username.
### Externally Participating in OASIS Server via curl (Complete Methods)
The OASIS Server (port 51202), in addition to being called by MCP tools, also supports direct curl operations for external scripts or debugging. All endpoints use `user_id` parameter for user isolation.
#### 1. Expert Management
```bash
# List all experts (public + user custom)
curl 'http://127.0.0.1:51202/experts?user_id=xinyuan'
# Create custom expert
curl -X POST 'http://127.0.0.1:51202/experts/user' \
-H 'Content-Type: application/json' \
-d '{"user_id":"xinyuan","name":"Product Manager","tag":"pm","persona":"You are an experienced product manager skilled in requirements analysis and product planning","temperature":0.7}'
# Update custom expert
curl -X PUT 'http://127.0.0.1:51202/experts/user/pm' \
-H 'Content-Type: application/json' \
-d '{"user_id":"xinyuan","persona":"Updated expert description"}'
# Delete custom expert
curl -X DELETE 'http://127.0.0.1:51202/experts/user/pm?user_id=xinyuan'
```
#### 2. Session Management
```bash
# List OASIS-managed expert sessions (sessions containing #oasis#)
curl 'http://127.0.0.1:51202/sessions/oasis?user_id=xinyuan'
```
#### 3. Workflow Management
```bash
# List user's saved workflows
curl 'http://127.0.0.1:51202/workflows?user_id=xinyuan'
# Save workflow (auto-generate layout)
curl -X POST 'http://127.0.0.1:51202/workflows' \
-H 'Content-Type: application/json' \
-d '{"user_id":"xinyuan","name":"trio_discussion","schedule_yaml":"version:1\nplan:\n - expert: \"creative#temp#1\"","description":"Trio discussion","save_layout":true}'
```
#### 4. Layout Generation
```bash
# Generate layout from YAML
curl -X POST 'http://127.0.0.1:51202/layouts/from-yaml' \
-H 'Content-Type: application/json' \
-d '{"user_id":"xinyuan","yaml_source":"version:1\nplan:\n - expert: \"creative#temp#1\"","layout_name":"trio_layout"}'
```
#### 5. Discussion/Execution
```bash
# Create discussion topic (synchronous, wait for conclusion)
curl -X POST 'http://127.0.0.1:51202/topics' \
-H 'Content-Type: application/json' \
-d '{"user_id":"xinyuan","question":"Discussion topic","max_rounds":3,"schedule_yaml":"version:1\nplan:\n - expert: \"creative#temp#1\"","discussion":true}'
# Create discussion topic (async, returns topic_id)
curl -X POST 'http://127.0.0.1:51202/topics' \
-H 'Content-Type: application/json' \
-d '{"user_id":"xinyuan","question":"Discussion topic","max_rounds":3,"schedule_yaml":"version:1\nplan:\n - expert: \"creative#temp#1\"","discussion":true,"callback_url":"http://127.0.0.1:51200/system_trigger","callback_session_id":"my-session"}'
# View discussion details
curl 'http://127.0.0.1:51202/topics/{topic_id}?user_id=xinyuan'
# Get discussion conclusion (blocking wait)
curl 'http://127.0.0.1:51202/topics/{topic_id}/conclusion?user_id=xinyuan&timeout=300'
# Cancel discussion
curl -X DELETE 'http://127.0.0.1:51202/topics/{topic_id}?user_id=xinyuan'
# List all discussion topics
curl 'http://127.0.0.1:51202/topics?user_id=xinyuan'
```
#### 6. Real-time Stream
```bash
# SSE real-time update stream (discussion mode)
curl 'http://127.0.0.1:51202/topics/{topic_id}/stream?user_id=xinyuan'
```
**Storage Locations:**
- Workflows (YAML): `data/user_files/{user}/oasis/yaml/{file}.yaml` (canvas layouts are converted from YAML in real-time, no longer stored as separate layout JSON)
- User custom experts: `data/oasis_user_experts/{user}.json`
- Discussion records: `data/oasis_topics/{user}/{topic_id}.json`
**Note:** These endpoints share the same backend implementation as MCP tools `list_oasis_experts`, `add_oasis_expert`, `update_oasis_expert`, `delete_oasis_expert`, `list_oasis_sessions`, `set_oasis_workflow`, `list_oasis_workflows`, `yaml_to_layout`, `post_to_oasis`, `check_oasis_discussion`, `cancel_oasis_discussion`, `list_oasis_topics`, ensuring consistent behavior.
## Example Configuration Reference
Below is an actual running configuration example (sensitive info redacted):
```bash
bash selfskill/scripts/run.sh configure --init
bash selfskill/scripts/run.sh configure --batch \
LLM_API_KEY=sk-xxxxxxxxxxxxxxxxxxxxxxxxxxxx4c74 \
LLM_BASE_URL=https://deepseek.com \
LLM_MODEL=deepseek-chat \
LLM_VISION_SUPPORT=true \
TTS_MODEL=gemini-2.5-flash-preview-tts \
TTS_VOICE=charon \
PORT_AGENT=51200 \
PORT_SCHEDULER=51201 \
PORT_OASIS=51202 \
PORT_FRONTEND=51209 \
PORT_BARK=58010 \
OPENCLAW_API_URL=http://127.0.0.1:18789/v1/chat/completions \
OPENAI_STANDARD_MODE=false
bash selfskill/scripts/run.sh add-user system <your-password>
```
Output after `configure --show`:
```
PORT_SCHEDULER=51201
PORT_AGENT=51200
PORT_FRONTEND=51209
PORT_OASIS=51202
OASIS_BASE_URL=http://127.0.0.1:51202
PORT_BARK=58010
INTERNAL_TOKEN=f1aa****57e7 # Auto-generated, do not leak
LLM_API_KEY=sk-7****4c74
LLM_BASE_URL=https://deepseek.com
LLM_MODEL=deepseek-chat
LLM_VISION_SUPPORT=true
TTS_MODEL=gemini-2.5-flash-preview-tts
TTS_VOICE=charon
OPENAI_STANDARD_MODE=false
```
> Note: `INTERNAL_TOKEN` is auto-generated on first startup; `PUBLIC_DOMAIN` / `BARK_PUBLIC_URL` are auto-written by the tunnel; no manual configuration needed.
## Typical Usage Flow
```bash
cd /home/avalon/TeamClaw
# First-time configuration
bash selfskill/scripts/run.sh setup
bash selfskill/scripts/run.sh configure --init
bash selfskill/scripts/run.sh configure --batch LLM_API_KEY=sk-xxx LLM_BASE_URL=https://api.deepseek.com LLM_MODEL=deepseek-chat
bash selfskill/scripts/run.sh add-user system MyPass123
# Start
bash selfskill/scripts/run.sh start
# Call API
curl -X POST http://127.0.0.1:51200/v1/chat/completions \
-H "Content-Type: application/json" \
-H "Authorization: Bearer system:MyPass123" \
-d '{"model":"mini-timebot","messages":[{"role":"user","content":"Hello"}],"stream":false,"session_id":"default"}'
# Stop
bash selfskill/scripts/run.sh stop
```
## Important Notes
- All skill scripts are in `selfskill/scripts/`, not affecting original project functionality
- Process management via PID files, `start` supports idempotent calls
- Do not leak `INTERNAL_TOKEN`
- Log path: `logs/launcher.log`
- Be sure to tell users how to open the visual interface and how to log in to the account for discussions
- The OpenClaw session file path can be inferred from the example path and current skill path. If you are an OpenClaw agent, be sure to complete the full OpenClaw-related configuration
---
---
# TeamClaw — Agent 子系统 Skill（中文版）
## 简介
TeamClaw 是一个类 OpenClaw 的多 Agent 子平台，拥有类似 OpenClaw 的内置轻量级 Agent，附带 computer use 功能以及社交软件（如 Telegram）集成。它可以在不阻塞主体 agent 的情况下独立运行，也可以由 OpenClaw agent 直接控制内置的 OASIS 协作平台进行多 Agent 编排。同时支持通过 Cloudflare 将前端暴露到公网，便于在手机页面等远程设备上展开可视化多 Agent 工作流编程。
TeamClaw 是一个多功能 AI Agent 服务，提供：
- **对话 Agent**：基于 LangGraph 的多工具 AI 助手，支持流式/非流式对话
- **OASIS 论坛**：多专家并行讨论/执行引擎，可编排多个 Agent 协作
- **定时调度**：基于 APScheduler 的任务调度中心
- **Bark 推送**：移动端推送通知
- **前端 Web UI**：完整的聊天界面
## Skill 脚本
所有脚本位于 `selfskill/scripts/`，统一通过 `run.sh` 入口调用，**全部非交互式**。
```
selfskill/scripts/
├── run.sh # 主入口（start/stop/status/setup/add-user/configure）
├── adduser.py # 非交互式用户创建
└── configure.py # 非交互式 .env 配置管理
```
## 快速启动
所有命令在项目根目录下执行。
### 1. 首次部署
```bash
# 安装依赖
bash selfskill/scripts/run.sh setup
# 初始化配置文件
bash selfskill/scripts/run.sh configure --init
# 配置 LLM（必填）
bash selfskill/scripts/run.sh configure --batch \
LLM_API_KEY=sk-your-key \
LLM_BASE_URL=https://api.deepseek.com \
LLM_MODEL=deepseek-chat
# 创建用户
bash selfskill/scripts/run.sh add-user system MySecurePass123
```
### 2. 启动/停止/状态
```bash
bash selfskill/scripts/run.sh start # 后台启动
bash selfskill/scripts/run.sh status # 检查状态
bash selfskill/scripts/run.sh stop # 停止服务
```
### 3. 配置管理
```bash
# 查看当前配置（敏感值脱敏）
bash selfskill/scripts/run.sh configure --show
# 设置单项
bash selfskill/scripts/run.sh configure PORT_AGENT 51200
# 批量设置
bash selfskill/scripts/run.sh configure --batch TTS_MODEL=gemini-2.5-flash-preview-tts TTS_VOICE=charon
```
## 可配置项
| 配置项 | 说明 | 默认值 |
|--------|------|--------|
| `LLM_API_KEY` | LLM API 密钥（**必填**） | — |
| `LLM_BASE_URL` | LLM API 地址 | `https://api.deepseek.com` |
| `LLM_MODEL` | 模型名称 | `deepseek-chat` |
| `LLM_PROVIDER` | 厂商（google/anthropic/deepseek/openai，可自动推断） | 自动 |
| `LLM_VISION_SUPPORT` | 是否支持图片（可自动推断） | 自动 |
| `PORT_AGENT` | Agent 主服务端口 | `51200` |
| `PORT_SCHEDULER` | 定时调度端口 | `51201` |
| `PORT_OASIS` | OASIS 论坛端口 | `51202` |
| `PORT_FRONTEND` | Web UI 端口 | `51209` |
| `PORT_BARK` | Bark 推送端口 | `58010` |
| `TTS_MODEL` | TTS 模型（可选） | — |
| `TTS_VOICE` | TTS 声音（可选） | — |
| `OPENCLAW_API_URL` | OpenClaw 后端服务地址（完整路径，含 `/v1/chat/completions`） | `http://127.0.0.1:18789/v1/chat/completions` |
| `OPENCLAW_API_KEY` | OpenClaw 后端服务的 API Key（可选） | — |
| `OPENCLAW_SESSIONS_FILE` | OpenClaw sessions.json 文件的绝对路径（**使用 OpenClaw 时必须配置**） | `/projects/.moltbot/agents/main/sessions/sessions.json` |
| `INTERNAL_TOKEN` | 内部通信密钥（自动生成） | 自动 |
## 端口与服务
| 端口 | 服务 |
|------|------|
| 51200 | AI Agent 主服务 |
| 51201 | 定时调度 |
| 51202 | OASIS 论坛 |
| 51209 | Web UI |
## API 认证
### 方式 1：用户认证
```
Authorization: Bearer <user_id>:<password>
```
### 方式 2：内部 Token（服务间调用，推荐）
```
Authorization: Bearer <INTERNAL_TOKEN>:<user_id>
```
`INTERNAL_TOKEN` 首次启动自动生成，可通过 `configure --show-raw` 查看。
## 核心 API
**Base URL**: `http://127.0.0.1:51200`
### 对话（OpenAI 兼容）
```
POST /v1/chat/completions
Authorization: Bearer <token>
{"model":"mini-timebot","messages":[{"role":"user","content":"你好"}],"stream":true,"session_id":"my-session"}
```
### 系统触发（内部调用）
```
POST /system_trigger
X-Internal-Token: <INTERNAL_TOKEN>
{"user_id":"system","text":"请执行某任务","session_id":"task-001"}
```
### 终止会话
```
POST /cancel
{"user_id":"<user_id>","session_id":"<session_id>"}
```
## OASIS 四种运行模式（默认：讨论模式）
> 这里的"四个模式"是两组正交开关：
> - **讨论 vs 执行**：决定专家输出是"论坛式讨论/投票"，还是"工作流式执行/交付物"。
> - **同步 vs 脱离（detach）**：决定调用方是否阻塞等待结果。
### 1) 讨论模式 vs 执行模式
**讨论模式（discussion=true，默认）**
- 用途：多专家给出不同观点、利弊分析、争议点澄清，并可形成共识。
- 适用：方案评审、技术路线选择、需要"为什么"的问题。
**执行模式（discussion=false）**
- 用途：把 OASIS 当作编排器按计划顺序/并行完成任务，强调直接产出（代码/脚本/清单/定稿方案）。
- 适用：目标明确的交付任务，且不需要展开辩论。
### 2) 同步模式 vs 脱离模式（detach）
**同步（detach=false，默认）**
- 行为：调用 `post_to_oasis` 后等待完成，直接返回最终结果。
- 适用：任务较快、需要立刻拿到产物并继续迭代。
**脱离（detach=true）**
- 行为：立即返回 `topic_id`，后台继续运行/讨论；之后用 `check_oasis_discussion(topic_id)` 追踪进度与结果。
- 适用：多轮/多专家/耗时长/包含工具调用，不希望阻塞当前会话。
### 3) 自动选择规则（建议默认策略）
在未明确指定时，建议采用以下默认策略：
1. **默认 = 讨论 + 同步**
- `discussion=true`
- `detach=false`
2. 出现以下需求信号时，切换到 **执行模式**：
- "直接给最终版 / 可复制粘贴 / 可执行脚本 / 只要结论不要讨论"
- "按步骤生成 SOP / 清单 / 表格并定稿"
3. 出现以下需求信号时，切换到 **脱离模式**：
- "后台跑 / 我一会儿再看 / 先给 topic_id"
- 多专家并行、多轮（`max_rounds` 大）、或预计耗时长
### 4) 四种组合速查
| 组合 | 参数 | 返回 | 适用 |
|---|---|---|---|
| 讨论 + 同步（默认） | discussion=true, detach=false | 当场看到讨论与结论 | 决策/评审/收集观点 |
| 讨论 + 脱离 | discussion=true, detach=true | topic_id，稍后查 | 长讨论/多轮 |
| 执行 + 同步 | discussion=false, detach=false | 直接交付物 | 生成代码/方案/清单 |
| 执行 + 脱离 | discussion=false, detach=true | topic_id，稍后查 | 长执行/复杂流水线 |
## OASIS 四类智能体
OASIS 支持 **四种类型的智能体**，通过 `schedule_yaml` 中专家的 `name` 格式区分：
| # | 类型 | Name 格式 | 引擎类 | 说明 |
|---|------|-----------|--------|------|
| 1 | **Direct LLM** | `tag#temp#N` | `ExpertAgent` | 无状态单次 LLM 调用。每轮读取所有帖子 → 一次 LLM 调用 → 发布 + 投票。无跨轮记忆。`tag` 映射到预设专家名/人设，`N` 是实例编号（同一专家可多副本）。 |
| 2 | **Oasis Session** | `tag#oasis#id` | `SessionExpert` (oasis) | OASIS 管理的有状态 bot session。`tag` 映射到预设专家，首轮注入人设为 system prompt。Bot 跨轮保留对话记忆（增量上下文）。`id` 可为任意字符串，新 ID 首次使用时自动创建 session。 |
| 3 | **Regular Agent** | `Title#session_id` | `SessionExpert` (regular) | 连接到已有的 agent session（如 `助手#default`、`Coder#my-project`）。不注入身份——session 自身的 system prompt 定义 agent。适合将个人 bot session 带入讨论。 |
| 4 | **External API** | `tag#ext#id` | `ExternalExpert` | 直接调用任意 OpenAI 兼容外部 API（DeepSeek、GPT-4、Ollama、另一个 TeamClaw 实例等）。不经过本地 agent。外部服务假定有状态。支持通过 YAML `headers` 字段自定义请求头。 | 经典用途：和openclaw agent连接
### Session ID 格式
```
tag#temp#N → ExpertAgent (无状态, 直连LLM)
tag#oasis#<id> → SessionExpert (oasis管理, 有状态bot)
Title#session_id → SessionExpert (普通agent session)
tag#ext#<id> → ExternalExpert (外部API，如openclaw agent)
```
**特殊后缀：**
- 在任意 session 名末尾追加 `#new` 可强制创建**全新 session**（ID 替换为随机 UUID，确保不复用）：
- `creative#oasis#abc#new` → `#new` 被剥离，ID 替换为 UUID
- `助手#my-session#new` → 同样处理
**Oasis session 约定：**
- Oasis session 通过 `session_id` 中的 `#oasis#` 标识（如 `creative#oasis#ab12cd34`）
- 存储在普通 Agent checkpoint DB（`data/agent_memory.db`）中，无独立存储
- 首次使用时自动创建，无需预创建
- `tag` 部分映射到预设专家配置以查找人设
### YAML 示例
```yaml
version: 1
plan:
# Type 1: Direct LLM（无状态，快速）
- expert: "creative#temp#1"
- expert: "critical#temp#2"
# Type 2: Oasis session（有状态，有记忆）
- expert: "data#oasis#analysis01"
- expert: "synthesis#oasis#new#new" # 强制全新session
# Type 3: Regular agent session（你现有的bot）
- expert: "助手#default"
- expert: "Coder#my-project"
# Type 4: External API（DeepSeek, GPT-4等）
- expert: "deepseek#ext#ds1"
# Type 4: OpenClaw External API（本地 Agent 服务）
- expert: "coder#ext#oc1"
api_url: "http://127.0.0.1:23001/v1/chat/completions"
model: "agent:main:test1" # agent:<agent_name>:<session>，session 不存在时自动新建
# 并行执行
- parallel:
- expert: "creative#temp#1"
instruction: "从创新角度分析"
- expert: "critical#temp#2"
instruction: "从风险角度分析"
# 全员发言 + 手动注入
- all_experts: true
- manual:
author: "主持人"
content: "请聚焦可行性"
```
### External API (Type 4) 详细配置
Type 4 外部 agent 支持在 YAML 步骤中提供额外配置字段：
```yaml
version: 1
plan:
- expert: "分析师#ext#analyst"
api_url: "https://api.deepseek.com" # 必填：外部 API 的 base URL（自动补全为 /v1/chat/completions）
api_key: "sk-xxx" # 必填：API key → Authorization: Bearer <key>
model: "deepseek-chat" # 可选：模型名，默认 gpt-3.5-turbo
headers: # 可选：自定义 HTTP 请求头（key-value 字典）
X-Custom-Header: "value"
```
**配置字段说明：**
| 字段 | 必填 | 说明 |
|------|------|------|
| `api_url` | ✅ | 外部 API 地址，自动补全路径为 `/v1/chat/completions` |
| `api_key` | ❌ | 放到 `Authorization: Bearer <key>` header 中 |
| `model` | ❌ | 默认 `gpt-3.5-turbo` |
| `headers` | ❌ | 任意 key-value 字典，合并到 HTTP 请求头 |
**OpenClaw 专属配置：**
OpenClaw 是一个本地运行的 OpenAI 兼容 Agent 服务。在 `.env` 中设置好 OpenClaw 专属 endpoint 后，前端编排面板中拖入 OpenClaw 专家时会**自动填入** `api_url` 和 `api_key`，无需手动输入：
```bash
# 配置 OpenClaw endpoint 和 sessions 文件路径
bash selfskill/scripts/run.sh configure --batch \
OPENCLAW_SESSIONS_FILE=/projects/.moltbot/agents/main/sessions/sessions.json \
OPENCLAW_API_URL=http://127.0.0.1:18789/v1/chat/completions \
OPENCLAW_API_KEY=your-openclaw-key-if-needed
```
> **⚠️ 注意：**
> - `OPENCLAW_SESSIONS_FILE` 是使用 OpenClaw 功能的**前提条件**，必须指向 OpenClaw 的 `sessions.json` 文件绝对路径。未配置时前端编排面板不会加载 OpenClaw sessions。
> - `OPENCLAW_API_URL` 应填写**完整路径**（含 `/v1/chat/completions`），系统会自动剥离后缀生成 base URL 填入 YAML。YAML 中的 `api_url` 字段只需要 base URL（如 `http://127.0.0.1:18789`），引擎会自动补全路径。
> - 如果你的 OpenClaw 服务运行在非默认端口，请务必修改这些配置。
**OpenClaw 的 `model` 字段格式：**
```
agent:<agent_name>:<session_name>
```
- `agent_name`：OpenClaw 中的 agent 名称，通常为 `main`
- `session_name`：会话名称，如 `test1`、`default` 等。**可以填入不存在的 session 名来自动新建**
示例：
- `agent:main:default` — 使用 main agent 的 default session
- `agent:main:test1` — 使用 main agent 的 test1 session（不存在则新建）
- `agent:main:code-review` — 使用 main agent 的 code-review session
**请求头组装逻辑：**
最终发出的请求头 = `Content-Type: application/json` + `Authorization: Bearer <api_key>`（如有） + YAML `headers` 中自定义的所有键值对。
---
## 独立使用 OASIS Server
OASIS Server（端口 51202）可以**独立于 Agent 主服务使用**。外部脚本、其他服务、或手动 curl 均可直接操作 OASIS 的全部功能，无需通过 MCP 工具或 Agent 对话。
**独立使用场景：**
- 从外部脚本自动发起多专家讨论/执行
- 调试 workflow 编排
- 将 OASIS 作为微服务集成到其他系统
- 管理专家、会话、workflow 等资源
**前提条件：**
- OASIS 服务已启动（`bash selfskill/scripts/run.sh start` 会同时启动所有服务）
- 所有接口使用 `user_id` 参数进行用户隔离（无需 Authorization header）
**API 概览：**
| 功能 | 方法 | 路径 |
|------|------|------|
| 列出专家 | GET | `/experts?user_id=xxx` |
| 创建自定义专家 | POST | `/experts/user` |
| 更新/删除自定义专家 | PUT/DELETE | `/experts/user/{tag}` |
| 列出 oasis sessions | GET | `/sessions/oasis?user_id=xxx` |
| 保存 workflow | POST | `/workflows` |
| 列出 workflows | GET | `/workflows?user_id=xxx` |
| YAML → Layout | POST | `/layouts/from-yaml` |
| 创建讨论/执行 | POST | `/topics` |
| 查看讨论详情 | GET | `/topics/{topic_id}?user_id=xxx` |
| 获取结论（阻塞） | GET | `/topics/{topic_id}/conclusion?user_id=xxx&timeout=300` |
| SSE 实时流 | GET | `/topics/{topic_id}/stream?user_id=xxx` |
| 取消讨论 | DELETE | `/topics/{topic_id}?user_id=xxx` |
| 列出所有话题 | GET | `/topics?user_id=xxx` |
> 这些接口与 MCP 工具共享同一后端实现，行为完全一致。
---
### OASIS 讨论/执行
```
POST http://127.0.0.1:51202/topics
{"question":"讨论主题","user_id":"system","max_rounds":3,"discussion":true,"schedule_file":"...","schedule_yaml":"...","callback_url":"http://127.0.0.1:51200/system_trigger","callback_session_id":"my-session"}
```
注意优先使用schedule_yaml避免重复输入yaml，这是yaml工作流文件的绝对路径，一般在/XXXXX/TeamClaw/data/user_files/username下
### 外部 curl 参与 OASIS 服务器（完整方法）
OASIS 服务器（端口 51202）除了供 MCP 工具调用外，也支持直接 curl 操作，便于外部脚本或调试。所有接口均使用 `user_id` 参数进行用户隔离。
#### 1. 专家管理
```bash
# 列出所有专家（公共 + 用户自定义）
curl 'http://127.0.0.1:51202/experts?user_id=xinyuan'
# 创建自定义专家
curl -X POST 'http://127.0.0.1:51202/experts/user' \
-H 'Content-Type: application/json' \
-d '{"user_id":"xinyuan","name":"产品经理","tag":"pm","persona":"你是一个经验丰富的产品经理，擅长需求分析和产品规划","temperature":0.7}'
# 更新自定义专家
curl -X PUT 'http://127.0.0.1:51202/experts/user/pm' \
-H 'Content-Type: application/json' \
-d '{"user_id":"xinyuan","persona":"更新后的专家描述"}'
# 删除自定义专家
curl -X DELETE 'http://127.0.0.1:51202/experts/user/pm?user_id=xinyuan'
```
#### 2. 会话管理
```bash
# 列出 OASIS 管理的专家会话（含 #oasis# 的 session）
curl 'http://127.0.0.1:51202/sessions/oasis?user_id=xinyuan'
```
#### 3. Workflow 管理
```bash
# 列出用户保存的 workflows
curl 'http://127.0.0.1:51202/workflows?user_id=xinyuan'
# 保存 workflow（自动生成 layout）
curl -X POST 'http://127.0.0.1:51202/workflows' \
-H 'Content-Type: application/json' \
-d '{"user_id":"xinyuan","name":"trio_discussion","schedule_yaml":"version:1\nplan:\n - expert: \"creative#temp#1\"","description":"三人讨论","save_layout":true}'
```
#### 4. Layout 生成
```bash
# 从 YAML 生成 layout
curl -X POST 'http://127.0.0.1:51202/layouts/from-yaml' \
-H 'Content-Type: application/json' \
-d '{"user_id":"xinyuan","yaml_source":"version:1\nplan:\n - expert: \"creative#temp#1\"","layout_name":"trio_layout"}'
```
#### 5. 讨论/执行
```bash
# 创建讨论话题（同步等待结论）
curl -X POST 'http://127.0.0.1:51202/topics' \
-H 'Content-Type: application/json' \
-d '{"user_id":"xinyuan","question":"讨论主题","max_rounds":3,"schedule_yaml":"version:1\nplan:\n - expert: \"creative#temp#1\"","discussion":true}'
# 创建讨论话题（异步，返回 topic_id）
curl -X POST 'http://127.0.0.1:51202/topics' \
-H 'Content-Type: application/json' \
-d '{"user_id":"xinyuan","question":"讨论主题","max_rounds":3,"schedule_yaml":"version:1\nplan:\n - expert: \"creative#temp#1\"","discussion":true,"callback_url":"http://127.0.0.1:51200/system_trigger","callback_session_id":"my-session"}'
# 查看讨论详情
curl 'http://127.0.0.1:51202/topics/{topic_id}?user_id=xinyuan'
# 获取讨论结论（阻塞等待）
curl 'http://127.0.0.1:51202/topics/{topic_id}/conclusion?user_id=xinyuan&timeout=300'
# 取消讨论
curl -X DELETE 'http://127.0.0.1:51202/topics/{topic_id}?user_id=xinyuan'
# 列出所有讨论话题
curl 'http://127.0.0.1:51202/topics?user_id=xinyuan'
```
#### 6. 实时流
```bash
# SSE 实时更新流（讨论模式）
curl 'http://127.0.0.1:51202/topics/{topic_id}/stream?user_id=xinyuan'
```
**保存位置：**
- Workflows (YAML): `data/user_files/{user}/oasis/yaml/{file}.yaml`（画布布局从 YAML 实时转换，不再单独存储 layout JSON）
- 用户自定义专家: `data/oasis_user_experts/{user}.json`
- 讨论记录: `data/oasis_topics/{user}/{topic_id}.json`
**注意：** 这些接口与 MCP 工具 `list_oasis_experts`、`add_oasis_expert`、`update_oasis_expert`、`delete_oasis_expert`、`list_oasis_sessions`、`set_oasis_workflow`、`list_oasis_workflows`、`yaml_to_layout`、`post_to_oasis`、`check_oasis_discussion`、`cancel_oasis_discussion`、`list_oasis_topics` 共享同一后端实现，确保行为一致。
## 案例配置参考
以下是一份实际运行的配置示例（敏感信息已脱敏）：
```bash
bash selfskill/scripts/run.sh configure --init
bash selfskill/scripts/run.sh configure --batch \
LLM_API_KEY=sk-xxxxxxxxxxxxxxxxxxxxxxxxxxxx4c74 \
LLM_BASE_URL=https://deepseek.com \
LLM_MODEL=deepseek-chat \
LLM_VISION_SUPPORT=true \
TTS_MODEL=gemini-2.5-flash-preview-tts \
TTS_VOICE=charon \
PORT_AGENT=51200 \
PORT_SCHEDULER=51201 \
PORT_OASIS=51202 \
PORT_FRONTEND=51209 \
PORT_BARK=58010 \
OPENCLAW_API_URL=http://127.0.0.1:18789/v1/chat/completions \
OPENAI_STANDARD_MODE=false
bash selfskill/scripts/run.sh add-user system <your-password>
```
配置完成后 `configure --show` 输出：
```
PORT_SCHEDULER=51201
PORT_AGENT=51200
PORT_FRONTEND=51209
PORT_OASIS=51202
OASIS_BASE_URL=http://127.0.0.1:51202
PORT_BARK=58010
INTERNAL_TOKEN=f1aa****57e7 # 自动生成，勿泄露
LLM_API_KEY=sk-7****4c74
LLM_BASE_URL=https://deepseek.com
LLM_MODEL=deepseek-chat
LLM_VISION_SUPPORT=true
TTS_MODEL=gemini-2.5-flash-preview-tts
TTS_VOICE=charon
OPENAI_STANDARD_MODE=false
```
> 说明：`INTERNAL_TOKEN` 首次启动自动生成，`PUBLIC_DOMAIN` / `BARK_PUBLIC_URL` 由 tunnel 自动写入，无需手动配置。
## 典型使用流程
```bash
cd /home/avalon/TeamClaw
# 首次配置
bash selfskill/scripts/run.sh setup
bash selfskill/scripts/run.sh configure --init
bash selfskill/scripts/run.sh configure --batch LLM_API_KEY=sk-xxx LLM_BASE_URL=https://api.deepseek.com LLM_MODEL=deepseek-chat
bash selfskill/scripts/run.sh add-user system MyPass123
# 启动
bash selfskill/scripts/run.sh start
# 调用 API
curl -X POST http://127.0.0.1:51200/v1/chat/completions \
-H "Content-Type: application/json" \
-H "Authorization: Bearer system:MyPass123" \
-d '{"model":"mini-timebot","messages":[{"role":"user","content":"你好"}],"stream":false,"session_id":"default"}'
# 停止
bash selfskill/scripts/run.sh stop
```
## 注意事项
- 所有 skill 脚本位于 `selfskill/scripts/`，不影响项目原有功能
- 通过 PID 文件管理进程，`start` 支持幂等调用
- `INTERNAL_TOKEN` 勿泄露
- 日志路径: `logs/launcher.log`
- 一定要告诉用户如何开启可视化界面，以及如何登录到进行讨论等工作的账号
- openclaw session file路径可以综合案例路径和当前skill路径推理得到。假如你是openclaw agent，一定要人类完成完整的openclaw相关配置
don't have the plugin yet? install it then click "run inline in claude" again.