Skill flagged — suspicious patterns detected

ClawHub Security flagged this skill as suspicious. Review the scan results before using.

Ai Cost Optimizer

v1.0.0

AI 模型成本优化器 - 自动计算 API 费用,推荐最优模型,对比 DeepSeek/智谱/通义/GPT 成本。省钱必备工具。

0· 108·0 current·0 all-time

Install

OpenClaw Prompt Flow

Install with OpenClaw

Best for remote or guided setup. Copy the exact prompt, then paste it into OpenClaw for yang1002378395-cmyk/ai-cost-optimizer-cn.

Previewing Install & Setup.
Prompt PreviewInstall & Setup
Install the skill "Ai Cost Optimizer" (yang1002378395-cmyk/ai-cost-optimizer-cn) from ClawHub.
Skill page: https://clawhub.ai/yang1002378395-cmyk/ai-cost-optimizer-cn
Keep the work scoped to this skill only.
After install, inspect the skill metadata and help me finish setup.
Use only the metadata you can verify from ClawHub; do not invent missing requirements.
Ask before making any broader environment changes.

Command Line

CLI Commands

Use the direct CLI path if you want to install manually and keep every step visible.

OpenClaw CLI

Bare skill slug

openclaw skills install ai-cost-optimizer-cn

ClawHub CLI

Package manager switcher

npx clawhub@latest install ai-cost-optimizer-cn
Security Scan
VirusTotalVirusTotal
Benign
View report →
OpenClawOpenClaw
Suspicious
medium confidence
!
Purpose & Capability
SKILL.md advertises features beyond what's included (batch analysis, API integration, automated reports, Pro capabilities), and even shows usage of report.js, but no report.js is present. The shipped code implements local calculators and recommendations only; claims about automated API integration and '自动统计' are not implemented in the provided files.
Instruction Scope
Runtime instructions mainly run local node scripts (calculate.js, compare.js, recommend.js) which only do local arithmetic and interactive prompts — they do not read secrets or call remote endpoints. However SKILL.md instructs users to edit ~/.openclaw/config.json and suggests running an external installer (npx clawhub@latest install openclaw-cn-installer / setup-ai.js), which would affect agent configuration outside this skill and could execute remote code if run. Also SKILL.md references report.js (missing), so the documentation and instructions are out-of-sync with delivered code.
Install Mechanism
There is no install spec in the registry and the provided scripts are local Node files (no downloads). That is low-risk. The documentation suggests using npx to install other tooling (openclaw-cn-installer / clawhub), which would run code fetched from npm and is an external action the user should review before running.
Credentials
The skill does not request environment variables, credentials, or config paths programmatically. SKILL.md shows optional environment variables and a configuration file (~/.openclaw/config.json) that users can edit to change default model settings; these are reasonable for changing model selection but are not required by the code.
Persistence & Privilege
The skill does not request always:true, does not autonomously modify other skills, and the code does not persist credentials. However the documentation encourages editing the global OpenClaw config (~/.openclaw/config.json) and using an external installer to set models, which — if followed — would change agent-wide behavior. That is an out-of-band action the user should review carefully.
What to consider before installing
This package's scripts (calculate.js, compare.js, recommend.js) are local calculators and are coherent with the stated cost-comparison purpose. However: (1) SKILL.md advertises extra features (report generation, API integration, Pro automated stats) that are not present in the shipped files — report.js is referenced but missing. (2) The doc suggests running external installers via npx (clawhub / openclaw-cn-installer) and editing ~/.openclaw/config.json; those actions fetch and run third-party code and change global config, so review that installer's source before running. If you expect the Pro/API features, ask the author for the missing files (report.js, API integration) or a trustworthy distribution. Do not run npx installs or edit global config unless you trust the external package and understand the changes. If you want higher assurance, request the maintainer's repo URL or a signed release before installing.

Like a lobster shell, security has layers — review code before you run it.

chinesevk975dg7fcngcpjhzt2t3kymdjx83jtmzcostvk975dg7fcngcpjhzt2t3kymdjx83jtmzdeepseekvk975dg7fcngcpjhzt2t3kymdjx83jtmzlatestvk975dg7fcngcpjhzt2t3kymdjx83jtmzpricingvk975dg7fcngcpjhzt2t3kymdjx83jtmz
108downloads
0stars
1versions
Updated 1mo ago
v1.0.0
MIT-0

AI 成本优化器

智能对比 AI 模型 API 成本,推荐最优模型,帮你省 90%+ API 费用。

功能

  • 💰 成本计算 - 实时计算不同模型的 API 费用
  • 📊 价格对比 - 一键对比 10+ 主流模型
  • 🎯 智能推荐 - 根据场景推荐最优模型
  • 📈 费用预估 - 预估使用 10 万 tokens 花多少钱
  • 💡 省钱建议 - 告诉你如何优化成本

安装

npx clawhub@latest install ai-cost-optimizer

使用方法

1. 对比模型价格

node ~/.openclaw/skills/ai-cost-optimizer/compare.js

输出示例:

📊 AI 模型价格对比
┌─────────────┬──────────┬──────────┬─────────────┐
│ 模型        │ 输入     │ 输出     │ 每10万tokens │
├─────────────┼──────────┼──────────┼─────────────┤
│ DeepSeek V3 │ ¥0.27/M  │ ¥1.08/M  │ ¥1.35      │
│ 智谱 GLM-4  │ ¥0.10/M  │ ¥0.10/M  │ ¥0.50      │
│ 通义千问    │ ¥0.30/M  │ ¥0.60/M  │ ¥1.80      │
│ GPT-4o      │ ¥35/M    │ ¥105/M   │ ¥140       │
│ Claude 3.5  │ ¥17.50/M │ ¥52.50/M │ ¥70        │
└─────────────┴──────────┴──────────┴─────────────┘

💡 推荐选择:
   日常使用 → DeepSeek V3(省 99% 费用)
   中文内容 → 智谱 GLM-4(省 99.5% 费用)
   复杂推理 → Claude 3.5(省 50% 费用)

2. 计算使用费用

# 计算特定模型的费用
node ~/.openclaw/skills/ai-cost-optimizer/calculate.js deepseek 500000

# 输出:使用 50 万 tokens 花费 ¥6.75

# 交互式计算
node ~/.openclaw/skills/ai-cost-optimizer/calculate.js

3. 智能推荐

node ~/.openclaw/skills/ai-cost-optimizer/recommend.js

会询问你的使用场景,然后推荐最优模型:

❓ 主要用途?
  1. 日常对话/写作
  2. 编程/代码生成
  3. 中文内容创作
  4. 复杂推理/分析
  5. 多模态(图文)

❓ 每天大约调用次数?
  1. < 100 次
  2. 100-1000 次
  3. 1000-10000 次
  4. > 10000 次

💡 推荐:DeepSeek V3
   - 性价比最高
   - 预估每月费用:¥27(1000次/天)
   - 同等使用 GPT-4o:¥2800(省 99%)

4. 生成成本报告

node ~/.openclaw/skills/ai-cost-optimizer/report.js --model deepseek --tokens 1000000

生成详细成本分析报告:

  • 各模型费用对比
  • 省钱潜力分析
  • 使用建议

支持的模型

模型输入价格输出价格推荐场景
DeepSeek V3¥0.27/M¥1.08/M日常对话、编程、写作
DeepSeek Coder¥0.27/M¥1.08/M代码生成、代码审查
智谱 GLM-4¥0.10/M¥0.10/M中文内容创作
智谱 GLM-3-Turbo¥0.05/M¥0.05/M高频调用
通义千问 Plus¥0.30/M¥0.60/M多模态任务
通义千问 Turbo¥0.10/M¥0.10/M快速响应
GPT-4o¥35/M¥105/M复杂推理(企业级)
GPT-4o-mini¥3.50/M¥10.50/M中等任务
Claude 3.5 Sonnet¥17.50/M¥52.50/M长文本、代码
Claude 3.5 Haiku¥0.18/M¥0.54/M快速对话

省钱案例

案例 1:个人开发者

  • 使用场景:每天 AI 写代码 + 写博客
  • 原方案:GPT-4o,每月 100 万 tokens → ¥1400
  • 优化后:DeepSeek V3,每月 100 万 tokens → ¥13.5
  • 节省¥1386.5/月(99%)

案例 2:初创公司

  • 使用场景:客服 Bot,每月 5000 万 tokens
  • 原方案:GPT-4o-mini → ¥17500/月
  • 优化后:DeepSeek V3 → ¥675/月
  • 节省¥16825/月(96%)

案例 3:中文内容团队

  • 使用场景:写文章、营销文案,每月 3000 万 tokens
  • 原方案:GPT-4o → ¥42000/月
  • 优化后:智谱 GLM-4 + DeepSeek 组合 → ¥450/月
  • 节省¥41550/月(99%)

混合策略(最优解)

根据不同任务使用不同模型:

任务类型推荐模型原因
简单对话DeepSeek V3便宜
代码生成DeepSeek Coder代码能力强
中文写作智谱 GLM-4中文能力强
复杂推理Claude 3.5推理强
多模态通义千问 Plus图文处理

预期节省:50-99%(根据任务分布)

价格数据来源

注意:价格可能变动,以官方为准。

常见问题

Q: DeepSeek 和 GPT-4 差多少?

A: 日常任务(写作、编程、问答)几乎没差。复杂推理(数学证明、多步逻辑)GPT-4 略好,但贵 100 倍。

Q: 可以混用多个模型吗?

A: 可以!推荐策略:

  • 80% 用 DeepSeek/智谱(便宜)
  • 20% 用 GPT/Claude(复杂任务)
  • 预期节省 70-80%

Q: 如何切换模型?

A: 编辑 ~/.openclaw/config.json

{
  "model": {
    "provider": "deepseek",
    "model": "deepseek-chat"
  }
}

Q: 这些价格准确吗?

A: 官方价格,2026-03 更新。价格可能变动,请以官方为准。

配置 OpenClaw 使用不同模型

方式 1:配置文件

编辑 ~/.openclaw/config.json

{
  "model": {
    "provider": "deepseek",
    "model": "deepseek-chat"
  }
}

方式 2:环境变量

export DEFAULT_MODEL_PROVIDER=deepseek
export DEFAULT_MODEL_ID=deepseek-chat

方式 3:使用 openclaw-cn-installer

npx clawhub@latest install openclaw-cn-installer
node ~/.openclaw/skills/openclaw-cn-installer/setup-ai.js deepseek

相关资源


💬 Pro 版本(¥199)

免费版(当前)

  • 模型价格对比
  • 基础费用计算
  • 场景推荐

Pro 版(¥199)

  • ✅ 批量分析(100+ 模型)
  • ✅ 历史趋势图表
  • ✅ 自动化成本报告
  • ✅ API 集成(自动统计)
  • ✅ 团队协作功能
  • ✅ 1年更新支持

联系方式

  • QQ: 1002378395(中国用户)
  • Telegram: 待注册(海外用户)

添加 QQ 1002378395,发送"AI成本优化"获取 Pro 版信息


License

MIT(免费版) Pro 版:付费后可用

Comments

Loading comments...