部署zeroclaw+qwen+dingtalk+skills
ZeroClaw：轻量高效的AI助手工具 ZeroClaw是基于Rust重构的轻量级AI助手，相比资源消耗大的OpenClaw，具有无GC、高效运行的特点。支持22+主流AI提供商，包括OpenAI、Anthropic、Mistral等系列模型，以及OpenRouter和任何OpenAI兼容API。 安装配置简单： 克隆GitHub仓库后通过Cargo构建 提供交互式聊天、单次提问、Webhook
·
前言
最近 openclaw 太火了,但是资源消耗比较大，一般设备搞不了
zeroclaw 基于rust 重构的，资源消耗小，还无GC
ZeroClaw 默认支持 22+ AI 提供者，覆盖主流大模型：
OpenAI 系列（GPT-4o、GPT-4 Turbo 等）
Anthropic 系列（Claude 3.5 Sonnet 等）
Mistral 系列
OpenRouter（聚合网关）
任何 OpenAI 兼容 API
1：阿里百炼API生成
官方文档 https://bailian.console.aliyun.com/cn-beijing?tab=api#/api/
华北2（北京）：https://dashscope.aliyuncs.com/compatible-mode/v1
2：安装配置
rust安装的最新版（基于ubuntu22）
git clone https://github.com/zeroclaw-labs/zeroclaw.git #网络不好，可以手动下载
cd zeroclaw
cargo build --release --locked #--locked 不增加也没事
####cargo install --path . --force --locked 这步执行 因为配置了环境到 PATH里，
# 查看状态
zeroclaw status
# 交互式聊天
zeroclaw agent
# 单次提问
zeroclaw agent -m "用 Python 实现快速排序"
# 启动 Webhook 网关（默认 127.0.0.1:8080）
zeroclaw gateway
# 随机端口（更安全）
zeroclaw gateway --port 0
# 后台守护进程（长期运行）
zeroclaw daemon
config.tom 位置 ~/.zeroclaw/config.toml
可以在toml 配置 OR 用 zeroclaw onboard --interactive 配置
python3 version3.11+
api_key = "enc2:e*******************" #自行修改
default_provider = "custom:https://dashscope.aliyuncs.com/compatible-mode/v1"
default_model = "deepseek-v3.2"
#default_model = "qwen-max" qwen-plus qwen3.5-plus qwen-coder-plus qwen3.5-27b
default_temperature = 0.7
model_routes = []
embedding_routes = []
[observability]
backend = "log" #Observability backend: none, noop, log, prometheus, otel, opentelemetry, or otlp
[autonomy]
level = "supervised"
workspace_only = true
allowed_commands = [
"git",
"npm",
"cargo",
"ls",
"cat",
"grep",
"find",
"echo",
"pwd",
"wc",
"head",
"tail",
"curl",
"mkdir",
"shell",
"jq",
"aws",
"python",
"python3",
]
forbidden_paths = [
"/etc",
"/root",
"/home",
"/usr",
"/bin",
"/sbin",
"/lib",
"/opt",
"/boot",
"/dev",
"/proc",
"/sys",
"/var",
"/tmp",
"~/.ssh",
"~/.gnupg",
"~/.aws",
"~/.config",
]
max_actions_per_hour = 20
max_cost_per_day_cents = 500
require_approval_for_medium_risk = true
block_high_risk_commands = true
auto_approve = [
"file_read",
"file_write",
"memory_recall",
]
always_ask = []
[runtime]
kind = "native"
[runtime.docker]
image = "alpine:3.20"
network = "none"
memory_limit_mb = 512
cpu_limit = 1.0
read_only_rootfs = true
mount_workspace = true
allowed_workspace_roots = []
[reliability]
provider_retries = 2
provider_backoff_ms = 500
fallback_providers = []
api_keys = []
channel_initial_backoff_secs = 2
channel_max_backoff_secs = 60
scheduler_poll_secs = 15
scheduler_retries = 2
[reliability.model_fallbacks]
[scheduler]
enabled = true
max_tasks = 64
max_concurrent = 4
[agent]
compact_context = false
max_tool_iterations = 10
max_history_messages = 50
parallel_tools = false
tool_dispatcher = "auto"
[query_classification]
enabled = false
rules = []
[heartbeat]
enabled = false
interval_minutes = 30
[cron]
enabled = true
max_run_history = 50
[channels_config]
cli = true
message_timeout_secs = 300
[channels_config.dingtalk]
client_id = "ding*****************" #自行修改
client_secret = "SL**********************************vTO" #自行修改
allowed_users = ["*"]
[memory]
backend = "sqlite"
auto_save = true
hygiene_enabled = true
archive_after_days = 7
purge_after_days = 30
conversation_retention_days = 30
embedding_provider = "none"
embedding_model = "text-embedding-3-small"
embedding_dimensions = 1536
vector_weight = 0.7
keyword_weight = 0.3
min_relevance_score = 0.4
embedding_cache_size = 10000
chunk_max_tokens = 512
response_cache_enabled = false
response_cache_ttl_minutes = 60
response_cache_max_entries = 5000
snapshot_enabled = false
snapshot_on_hygiene = false
auto_hydrate = true
[storage.provider.config]
provider = ""
schema = "public"
table = "memories"
[tunnel]
provider = "none"
[gateway]
port = 3000
host = "127.0.0.1"
require_pairing = false
allow_public_bind = false
paired_tokens = []
pair_rate_limit_per_minute = 10
webhook_rate_limit_per_minute = 60
trust_forwarded_headers = false
rate_limit_max_keys = 10000
idempotency_ttl_secs = 300
idempotency_max_keys = 10000
[composio]
enabled = false
entity_id = "default"
[secrets]
encrypt = true
[browser]
enabled = false
allowed_domains = []
backend = "agent_browser"
native_headless = true
native_webdriver_url = "http://127.0.0.1:9515"
[browser.computer_use]
endpoint = "http://127.0.0.1:8787/v1/actions"
timeout_ms = 15000
allow_remote_endpoint = false
window_allowlist = []
[http_request]
enabled = true
allowed_domains = ["60s.viki.moe","*.com","*.cn"]
#allowed_domains = ["com", "org", "net", "io", "dev", "ai", "cn", "vip", "info", "tech", "app", "co", "me"]
max_response_size = 1000000
timeout_secs = 30
[multimodal]
max_images = 4
max_image_size_mb = 5
allow_remote_fetch = false
[web_search]
enabled = true
provider = "duckduckgo"
max_results = 5
timeout_secs = 15
[proxy]
enabled = false
no_proxy = []
scope = "zeroclaw"
services = []
[identity]
format = "openclaw"
[cost]
enabled = false
daily_limit_usd = 10.0
monthly_limit_usd = 100.0
warn_at_percent = 80
allow_override = false
[cost.prices."openai/gpt-4o-mini"]
input = 0.15
output = 0.6
[cost.prices."google/gemini-2.0-flash"]
input = 0.1
output = 0.4
[cost.prices."openai/o1-preview"]
input = 15.0
output = 60.0
[cost.prices."anthropic/claude-opus-4-20250514"]
input = 15.0
output = 75.0
[cost.prices."google/gemini-1.5-pro"]
input = 1.25
output = 5.0
[cost.prices."anthropic/claude-sonnet-4-20250514"]
input = 3.0
output = 15.0
[cost.prices."anthropic/claude-3.5-sonnet"]
input = 3.0
output = 15.0
[cost.prices."openai/gpt-4o"]
input = 5.0
output = 15.0
[cost.prices."anthropic/claude-3-haiku"]
input = 0.25
output = 1.25
[peripherals]
enabled = false
boards = []
[agents]
[hardware]
enabled = false
transport = "None"
baud_rate = 115200
workspace_datasheets = false
具体有哪些配置选项可以参考源码
zeroclaw\src\config\schema.rs
/// Top-level ZeroClaw configuration, loaded from `config.toml`.
///
/// Resolution order: `ZEROCLAW_WORKSPACE` env → `active_workspace.toml` marker → `~/.zeroclaw/config.toml`.
#[derive(Debug, Clone, Serialize, Deserialize, JsonSchema)]
pub struct Config {
/// Workspace directory - computed from home, not serialized
#[serde(skip)]
pub workspace_dir: PathBuf,
/// Path to config.toml - computed from home, not serialized
#[serde(skip)]
pub config_path: PathBuf,
/// API key for the selected provider. Overridden by `ZEROCLAW_API_KEY` or `API_KEY` env vars.
pub api_key: Option<String>,
/// Base URL override for provider API (e.g. "http://10.0.0.1:11434" for remote Ollama)
pub api_url: Option<String>,
/// Default provider ID or alias (e.g. `"openrouter"`, `"ollama"`, `"anthropic"`). Default: `"openrouter"`.
pub default_provider: Option<String>,
/// Default model routed through the selected provider (e.g. `"anthropic/claude-sonnet-4-6"`).
pub default_model: Option<String>,
/// Default model temperature (0.0–2.0). Default: `0.7`.
pub default_temperature: f64,
/// Observability backend configuration (`[observability]`).
#[serde(default)]
pub observability: ObservabilityConfig,
/// Autonomy and security policy configuration (`[autonomy]`).
#[serde(default)]
pub autonomy: AutonomyConfig,
/// Runtime adapter configuration (`[runtime]`). Controls native vs Docker execution.
#[serde(default)]
pub runtime: RuntimeConfig,
/// Reliability settings: retries, fallback providers, backoff (`[reliability]`).
#[serde(default)]
pub reliability: ReliabilityConfig,
/// Scheduler configuration for periodic task execution (`[scheduler]`).
#[serde(default)]
pub scheduler: SchedulerConfig,
/// Agent orchestration settings (`[agent]`).
#[serde(default)]
pub agent: AgentConfig,
/// Model routing rules — route `hint:<name>` to specific provider+model combos.
#[serde(default)]
pub model_routes: Vec<ModelRouteConfig>,
/// Embedding routing rules — route `hint:<name>` to specific provider+model combos.
#[serde(default)]
pub embedding_routes: Vec<EmbeddingRouteConfig>,
/// Automatic query classification — maps user messages to model hints.
#[serde(default)]
pub query_classification: QueryClassificationConfig,
/// Heartbeat configuration for periodic health pings (`[heartbeat]`).
#[serde(default)]
pub heartbeat: HeartbeatConfig,
/// Cron job configuration (`[cron]`).
#[serde(default)]
pub cron: CronConfig,
/// Channel configurations: Telegram, Discord, Slack, etc. (`[channels_config]`).
#[serde(default)]
pub channels_config: ChannelsConfig,
/// Memory backend configuration: sqlite, markdown, embeddings (`[memory]`).
#[serde(default)]
pub memory: MemoryConfig,
/// Persistent storage provider configuration (`[storage]`).
#[serde(default)]
pub storage: StorageConfig,
/// Tunnel configuration for exposing the gateway publicly (`[tunnel]`).
#[serde(default)]
pub tunnel: TunnelConfig,
/// Gateway server configuration: host, port, pairing, rate limits (`[gateway]`).
#[serde(default)]
pub gateway: GatewayConfig,
/// Composio managed OAuth tools integration (`[composio]`).
#[serde(default)]
pub composio: ComposioConfig,
/// Secrets encryption configuration (`[secrets]`).
#[serde(default)]
pub secrets: SecretsConfig,
/// Browser automation configuration (`[browser]`).
#[serde(default)]
pub browser: BrowserConfig,
/// HTTP request tool configuration (`[http_request]`).
#[serde(default)]
pub http_request: HttpRequestConfig,
/// Multimodal (image) handling configuration (`[multimodal]`).
#[serde(default)]
pub multimodal: MultimodalConfig,
/// Web search tool configuration (`[web_search]`).
#[serde(default)]
pub web_search: WebSearchConfig,
/// Proxy configuration for outbound HTTP/HTTPS/SOCKS5 traffic (`[proxy]`).
#[serde(default)]
pub proxy: ProxyConfig,
/// Identity format configuration: OpenClaw or AIEOS (`[identity]`).
#[serde(default)]
pub identity: IdentityConfig,
/// Cost tracking and budget enforcement configuration (`[cost]`).
#[serde(default)]
pub cost: CostConfig,
/// Peripheral board configuration for hardware integration (`[peripherals]`).
#[serde(default)]
pub peripherals: PeripheralsConfig,
/// Delegate agent configurations for multi-agent workflows.
#[serde(default)]
pub agents: HashMap<String, DelegateAgentConfig>,
/// Hardware configuration (wizard-driven physical world setup).
#[serde(default)]
pub hardware: HardwareConfig,
}
3：基本功能测试运行
4:接入百炼大模型
密钥管理
模型211个，每个有100W token ，先开启再用
5：接入dingtalk
https://open-dev.dingtalk.com/
创建应用
应用凭证与基础信息
6:增加skills
https://skills.sh/resend/resend-skills/send-email
命令增加 npx skills add https://github.com/resend/resend-skills --skill send-email
https://github.com/vikiboss/60s-skills
可以 npx skills add https://github.com/vikiboss/60s-skills --skill daily-news-60s
再拷贝到skills目录下
或 手动下载 自行拷贝进去
send-email 有问题，发送不了，所以这里做了修改
SKILL.md 修改后内容如下
---
name: send-email
description: Send emails via python3 script,not need SMTP configuration.
metadata: {"zeroclaw":{"emoji":"📧","requires":{"anyBins":["python3"]}}}
---
# Send Email
Send emails via the Python script,not need SMTP configuration. **Do not read** any config file (e.g. `~/.zeroclaw/config.toml` or `workspace/config.toml`) — that would expose credentials in tool output. Just run the script; env is injected automatically. Do not use ~/.msmtprc.
## Agent instructions
1. **Credentials**:when the script runs — do not use the read tool on `~/.zeroclaw/config.toml` or `workspace/config.toml` (exposes secrets). If the skill is enabled, assume env is configured; do not ask the user for passwords. Do not use ~/.msmtprc.
2. **Send mail**: Run the script under **workspace** (do not use the path under node_modules):
---bash
python3 ~/.zeroclaw/workspace/skills/send-email/send_email.py "recipient" "Subject" "Body"
---
3. **Attachment**: `python3 ~/.zeroclaw/workspace/skills/send-email/send_email.py "recipient" "Subject" "Body" "/path/to/file.pdf"`
## Usage examples
---bash
python3 ~/.zeroclaw/workspace/skills/send-email/send_email.py 'recipient@example.com' 'Subject' 'Body text'
python3 ~/.zeroclaw/workspace/skills/send-email/send_email.py 'recipient@example.com' 'Subject' 'Body' '/path/to/file.pdf'
---
## SMTP reference
- 163: `smtp.163.com:465`, requires authorization code (not login password)
- Gmail: `smtp.gmail.com:587`, requires App Password
- QQ: `smtp.qq.com:465`, requires authorization code
## Troubleshooting
- Authentication failed: Check that `EMAIL_SMTP_PASSWORD` is the authorization code or App Password.
- Connection failed: Check `EMAIL_SMTP_SERVER` and `EMAIL_SMTP_PORT`.
send_email.py 也修改了
#!/usr/bin/env python3
# -*- coding: utf-8 -*-
"""
邮件发送脚本 - OpenClaw Skill
支持命令行参数和环境变量配置
"""
import smtplib
import os
import sys
import tomllib
from email.mime.text import MIMEText
from email.mime.multipart import MIMEMultipart
from email.utils import formataddr
# ========== 配置区域 - 仅从环境变量读取，无默认值 ==========
'''
SMTP_SERVER = os.getenv("EMAIL_SMTP_SERVER")
_port = os.getenv("EMAIL_SMTP_PORT")
SMTP_PORT = int(_port) if _port else None
SENDER_EMAIL = os.getenv("EMAIL_SENDER")
AUTHORIZATION_CODE = os.getenv("EMAIL_SMTP_PASSWORD")
'''
'''
SMTP_SERVER="smtp.163.com"
SMTP_PORT=465
SENDER_EMAIL="*****@163.com"
AUTHORIZATION_CODE="*******"
USE_TLS = (os.getenv("EMAIL_USE_TLS") or "false").lower() == "true"
print("#######################")
#print("SMTP_SERVER={SMTP_SERVER}")
'''
# ===============================================
global_scope = {} # 全局作用域的字典模拟
def read_email_env():
file_path = os.path.expanduser("~/.zeroclaw/email.env")
with open(file_path, 'rb') as f:
config = tomllib.load(f)
if config['email']['EMAIL_SMTP_SERVER'] :
global_scope['EMAIL_SMTP_SERVER']=config['email']['EMAIL_SMTP_SERVER']
if config['email']['EMAIL_SMTP_PORT'] :
global_scope['EMAIL_SMTP_PORT']=config['email']['EMAIL_SMTP_PORT']
if config['email']['EMAIL_SENDER'] :
global_scope['EMAIL_SENDER']=config['email']['EMAIL_SENDER']
if config['email']['EMAIL_SMTP_PASSWORD'] :
global_scope['EMAIL_SMTP_PASSWORD']=config['email']['EMAIL_SMTP_PASSWORD']
print("###########################################" )
print( global_scope['EMAIL_SMTP_SERVER'])
print( global_scope['EMAIL_SMTP_PORT'])
print( global_scope['EMAIL_SENDER'])
print( global_scope['EMAIL_SMTP_PASSWORD'])
# =================================================
def get_value_by_key(stringkey):
if global_scope.get(stringkey):
return global_scope[stringkey]
else:
return ""
def send_email(to_email, subject, content, attachment_path=None):
"""发送邮件
Args:
to_email: 收件人邮箱
subject: 邮件主题
content: 邮件正文
attachment_path: 可选附件路径
Returns:
bool: 发送是否成功
"""
SMTP_SERVER = get_value_by_key("EMAIL_SMTP_SERVER")
_port = get_value_by_key("EMAIL_SMTP_PORT")
SMTP_PORT = int(_port) if _port else None
SENDER_EMAIL = get_value_by_key("EMAIL_SENDER")
AUTHORIZATION_CODE = get_value_by_key("EMAIL_SMTP_PASSWORD")
if not all([SMTP_SERVER, SMTP_PORT, SENDER_EMAIL, AUTHORIZATION_CODE]):
print("❌ 错误：请配置环境变量 EMAIL_SMTP_SERVER, EMAIL_SMTP_PORT, EMAIL_SENDER, EMAIL_SMTP_PASSWORD（在 ~/.openclaw/openclaw.json skills.entries.send-email.env）")
return False
try:
# 创建邮件
if attachment_path and os.path.exists(attachment_path):
msg = MIMEMultipart()
msg.attach(MIMEText(content, 'plain', 'utf-8'))
# 添加附件
from email.mime.base import MIMEBase
from email import encoders
with open(attachment_path, 'rb') as f:
part = MIMEBase('application', 'octet-stream')
part.set_payload(f.read())
encoders.encode_base64(part)
part.add_header(
'Content-Disposition',
f'attachment; filename= {os.path.basename(attachment_path)}'
)
msg.attach(part)
else:
msg = MIMEMultipart()
msg.attach(MIMEText(content, 'plain', 'utf-8'))
msg['From'] = formataddr(["OpenClaw", SENDER_EMAIL])
msg['To'] = to_email
msg['Subject'] = subject
# 连接 SMTP 服务器并发送
USE_TLS = False
if USE_TLS:
server = smtplib.SMTP(SMTP_SERVER, SMTP_PORT)
server.starttls()
else:
server = smtplib.SMTP_SSL(SMTP_SERVER, SMTP_PORT)
server.login(SENDER_EMAIL, AUTHORIZATION_CODE)
server.sendmail(SENDER_EMAIL, [to_email], msg.as_string())
server.quit()
print(f"✅ 邮件已成功发送到：{to_email}")
if attachment_path:
print(f" 附件：{os.path.basename(attachment_path)}")
return True
except Exception as e:
print(f"❌ 发送失败：{e}")
return False
if __name__ == "__main__":
if len(sys.argv) < 4:
print("用法：send_email.py <收件人> <主题> <正文> [附件路径]")
print("\n环境变量:")
print(" EMAIL_SMTP_SERVER SMTP 服务器（必填）")
print(" EMAIL_SMTP_PORT SMTP 端口（必填）")
print(" EMAIL_SENDER 发件人邮箱（必填）")
print(" EMAIL_SMTP_PASSWORD 授权码/密码（必填）")
print(" EMAIL_USE_TLS true 使用 TLS，否则 SSL（可选）")
sys.exit(1)
to_email = sys.argv[1]
subject = sys.argv[2]
content = sys.argv[3]
attachment = sys.argv[4] if len(sys.argv) > 4 else None
# print("SMTP_SERVER={SMTP_SERVER}")
read_email_env()
# sys.exit(1)
success = send_email(to_email, subject, content, attachment)
sys.exit(0 if success else 1)
email.env 内容如下
EMAIL_SMTP_PASSWORD= 不要写密码，是授权码
[email]
EMAIL_SMTP_SERVER="smtp.163.com"
EMAIL_SMTP_PORT="465"
EMAIL_SENDER="******@163.com"
EMAIL_SMTP_PASSWORD="**********"
7:如果觉得有用，麻烦点个赞，加个收藏
网上说zeroclaw 消耗token很大，qwen 有免费的额度，测试下来，下面这样，不到10次就用了100W TOKEN 了
skills 相关有待学习 ，skills 太多了
小龙虾开发者社区是 CSDN 旗下专注 OpenClaw 生态的官方阵地，聚焦技能开发、插件实践与部署教程，为开发者提供可直接落地的方案、工具与交流平台，助力高效构建与落地 AI 应用
更多推荐
23
0- 0
已为社区贡献1条内容
所有评论(0)