Dealpilot

v0.1.0

Cross-Platform Shopping Decision Agent / 全网购物决策官. Compares prices, quality, and risk across 淘宝/拼多多/京东/一号店/唯品会 and recommends the best platform for a given pr...

0· 103·0 current·0 all-time
byhaidong@harrylabsj

Install

OpenClaw Prompt Flow

Install with OpenClaw

Best for remote or guided setup. Copy the exact prompt, then paste it into OpenClaw for harrylabsj/dealpilot.

Previewing Install & Setup.
Prompt PreviewInstall & Setup
Install the skill "Dealpilot" (harrylabsj/dealpilot) from ClawHub.
Skill page: https://clawhub.ai/harrylabsj/dealpilot
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 dealpilot

ClawHub CLI

Package manager switcher

npx clawhub@latest install dealpilot
Security Scan
VirusTotalVirusTotal
Benign
View report →
OpenClawOpenClaw
Benign
high confidence
Purpose & Capability
Name/description (cross-platform shopping decision engine) match the included code and docs: engine, platform adapter stubs, normalize/decide/analyze scripts, and test-stub. Nothing in the repo asks for unrelated cloud credentials or system access.
Instruction Scope
SKILL.md and RUNNING.md limit runtime actions to local scripts (normalize/decide/format) and describe future integration with other platform skills. Example paths reference a local development path but there are no instructions to read unrelated system files or exfiltrate data. Current adapters are stubs returning mock data.
Install Mechanism
No install spec provided and no downloads or extract steps. The package is a local JS/TS code bundle intended to run in the agent environment; this is low-risk for installation mechanism.
Credentials
The skill declares no required env vars, binaries, or config paths. That is proportionate for the current stubbed implementation. Note: real platform adapters (future work) may legitimately require platform API keys—those should be requested only when adapter code is added and justified.
Persistence & Privilege
Skill is user-invocable and not always-enabled; it does not request permanent presence or modify other skills. There is no evidence it attempts to change system-wide settings or persist credentials.
Assessment
This skill is a coherent MVP scaffold that currently returns mock results and does not request credentials. If you install it now, it should be low-risk, but watch for future updates: real platform adapters will likely require API keys or third‑party integrations (Taobao/PDD/JD), and those additions should be reviewed to ensure requested credentials and network endpoints match the platform being integrated. Also verify any runtime network calls and new install steps before granting secrets or broad permissions.

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

latestvk978795h7frf028n2pxqze2qrs849c8b
103downloads
0stars
1versions
Updated 3w ago
v0.1.0
MIT-0

DealPilot / 全网购物决策官

你是全网购物决策官

你的任务不是罗列商品参数,而是在用户给出购物需求后,帮他在多个平台中找到最优解, 并给出可执行的购买建议。

产品定位

DealPilot 是一个跨平台购物比价决策引擎,覆盖:淘宝、拼多多、京东、一号店、唯品会。

核心价值:

  • 跨平台比价:同商品多平台实时价格对比
  • 风险评估:卖家信誉、售后保障、假货风险
  • 时机判断:当前是否是最佳购买时机
  • 推荐收敛:给出明确平台建议和理由

使用场景

用户可能会说:

  • "帮我看看这个产品在哪个平台买最划算"
  • "对比一下京东和拼多多买这个"
  • "我想买 X,百元以内,哪个平台靠谱"
  • "这个价格是不是好价,要不要等"
  • "淘宝和唯品会哪个更靠谱"

输入 schema(统一需求格式)

interface ShoppingRequest {
  // 商品信息
  product?: string;          // 商品名称/关键词
  productUrl?: string;       // 商品链接(可多平台)
  category?: string;         // 品类

  // 用户偏好
  budget?: PriceRange;       // 预算范围
  priorities?: string[];     // 优先级:["价格", "品质", "速度", "售后"]
  quantity?: number;         // 购买数量

  // 场景
  scenario?: "personal" | "gift" | "resale";  // 自用/送礼/转卖
  urgency?: "low" | "medium" | "high";        // 紧急程度

  // 平台范围(默认全部)
  platforms?: Platform[];

  // 约束
  mustHave?: string[];      // 必须有的特性/服务
  mustAvoid?: string[];     // 必须避免的
}

type Platform = "taobao" | "pdd" | "jd" | "yhd" | "vip";
type PriceRange = { min?: number; max?: number; };

输出 schema(统一决策报告)

interface DecisionReport {
  // 推荐决策
  recommendedPlatform: Platform;
  recommendedUrl?: string;
  recommendedPrice?: number;

  // 决策理由
  reasons: string[];           // 为什么要选这个平台

  // 风险提示
  risks: RiskItem[];          // 当前方案的风险点

  // 替代方案
  alternatives: AltPlatform[]; // 其他可选平台,排序

  // 时机建议
  timingAdvice: TimingAdvice;

  // 对比摘要(表格)
  comparison: PlatformSummary[];

  // 最终结论(可执行)
  conclusion: string;
}

interface RiskItem {
  level: "low" | "medium" | "high";
  description: string;
  mitigation?: string;  // 如何降低风险
}

interface AltPlatform {
  platform: Platform;
  score: number;        // 0-100
  price?: number;
  keyReason: string;
}

interface TimingAdvice {
  verdict: "buy_now" | "wait" | "compare_first";
  reason: string;
  waitDays?: number;    // 建议等多少天
  betterPeriod?: string; // "618", "双十一" 等
}

interface PlatformSummary {
  platform: Platform;
  price?: number;
  quality: "low" | "medium" | "high";
  shipping: "slow" | "medium" | "fast";
  afterSales: "weak" | "medium" | "strong";
  authenticity: "risky" | "medium" | "safe";
  score: number;
}

决策引擎流程

用户需求输入
    ↓
[需求解析] → 识别商品、预算、偏好、场景
    ↓
[平台搜索] → 各平台并行搜索(stub 阶段返回 mock 数据)
    ↓
[价格抓取] → 提取各平台到手价(含优惠/拼团等)
    ↓
[质量评估] → 店铺信誉、评论、售后标识
    ↓
[风险评分] → 假货/售后/物流风险量化
    ↓
[决策推荐] → 综合评分 + 时机判断
    ↓
[报告生成] → DecisionReport 输出

平台适配层

每个平台适配器需实现以下接口:

interface PlatformAdapter {
  platform: Platform;
  
  // 搜索商品
  search(query: string, options?: SearchOptions): Promise<SearchResult[]>;

  // 获取商品详情
  getProductDetail(url: string): Promise<ProductInfo>;

  // 计算最终到手价(含优惠/拼团/百亿补贴等)
  calcFinalPrice(product: ProductInfo): Promise<FinalPrice>;

  // 评估店铺/卖家风险
  assessSeller(url: string): Promise<SellerRisk>;

  // 平台特性判断
  getPlatformTraits(): PlatformTraits;
}

interface SearchOptions {
  category?: string;
  minPrice?: number;
  maxPrice?: number;
  sortBy?: "price" | "sales" | "rating";
}

interface FinalPrice {
  rawPrice: number;
  finalPrice: number;
  discountDesc: string[];   // ["百亿补贴-100", "拼团-30"]
  isSubsidized: boolean;    // 是否官方补贴
}

interface PlatformTraits {
  strength: string[];      // ["价格最低", "自营可信"]
  weakness: string[];       // ["物流较慢", "退换麻烦"]
  bestFor: string[];       // ["标准品", "电子产品"]
  worstFor: string[];       // ["生鲜", "奢侈品"]
}

平台优先级参考

维度淘宝拼多多京东一号店唯品会
价格最低中低
品质偏低
物流最快
售后
正品风险高最安全

当前状态

  • 搜索/比价:stub(返回 mock 数据)
  • 价格抓取:stub
  • 店铺评估:stub
  • 决策推荐:stub(固定返回 mock 推荐)
  • 时机判断:stub

下一步由平台组接入真实适配器。

相关 Skill

  • taobao - 淘宝平台适配
  • pdd-shopping - 拼多多平台适配
  • jingdong - 京东平台适配
  • shopping-advisor - 单商品决策逻辑

当前状态 (v0.1.0)

当前状态 (v0.1.0)

MVP 骨架版本 - 平台适配器为 stub 实现,返回 mock 数据。

已实现

  • ✅ 输入/输出 schema 定义
  • ✅ 决策引擎流程定义
  • ✅ 平台适配器接口定义
  • ✅ Mock 决策输出
  • ✅ 自测脚本 (scripts/test-stub.js)

待实现

  • 🔄 各平台真实搜索/价格抓取
  • 🔄 真实店铺风险评估
  • 🔄 实时价格/优惠信息
  • 🔄 接入现有平台 skill (taobao, pdd-shopping, jingdong)

使用说明

本地测试

cd /Users/jianghaidong/.openclaw/skills/dealpilot
node scripts/test-stub.js

在 OpenClaw 中调用

// 示例:调用 dealpilot 进行决策
const { normalizeRequest } = require('./scripts/normalize.js');
const { decide } = require('./scripts/decide.js');
const { formatReport } = require('./scripts/analyze.js');

async function runDealPilot(product, budget) {
  const request = normalizeRequest({ product, budget });
  const report = await decide(request);
  return formatReport(report);
}

// 示例调用
runDealPilot("蓝牙耳机", { max: 200 }).then(console.log);

平台适配器状态

平台适配器状态备注
淘宝TaobaoAdapterstub待接入 taobao skill
拼多多PddAdapterstub待接入 pdd-shopping skill
京东JdAdapterstub待接入 jingdong skill
一号店YhdAdapterstub待接入 yhd skill
唯品会VipAdapterstub待接入 vip skill

开发指南

目录结构

dealpilot/
├── SKILL.md                 # 技能定义
├── clawhub.json            # 技能元数据
├── package.json            # 依赖配置
├── README.md               # 项目说明
├── engine/                 # 决策引擎
│   ├── router.ts          # 路由层(决策入口)
│   └── types.ts           # 类型定义
├── platforms/              # 平台适配器
│   ├── base.ts            # 适配器基类
│   ├── taobao.ts          # 淘宝适配器
│   ├── pdd.ts             # 拼多多适配器
│   ├── jd.ts              # 京东适配器
│   ├── yhd.ts             # 一号店适配器
│   └── vip.ts             # 唯品会适配器
└── scripts/               # 工具脚本
    ├── normalize.js       # 需求解析
    ├── decide.js          # 决策调用
    ├── analyze.js         # 报告格式化
    └── test-stub.js       # 自测脚本

下一步开发

  1. 接入真实平台数据 - 将各平台适配器的 stub 方法替换为真实 API 调用
  2. 价格抓取 - 实现各平台实时价格获取
  3. 风险评估 - 接入店铺信誉、评论分析
  4. 时机判断 - 结合促销日历、历史价格
  5. 集成测试 - 端到端测试真实商品决策

相关 Skill

  • taobao - 淘宝平台搜索与详情
  • pdd-shopping - 拼多多百亿补贴/拼团策略
  • jingdong - 京东自营/物流分析
  • shopping-advisor - 单商品深度分析

Comments

Loading comments...