← 返回 Skills 市场
airs-guest

AIRS 具身智能产业研究 Skills

作者 airs-git · GitHub ↗ · v1.0.1 · MIT-0
cross-platform ✓ 安全检测通过
53
总下载
0
收藏
0
当前安装
2
版本数
在 OpenClaw 中安装
/install airs-embodied-intelligence-research-skills
功能描述
AIRS 具身智能产业研究 Skills。面向 embodied intelligence、robotics 和机器人产业研究,将企业主体确认、天眼查招投标/中标公告采集、第三方订单核查、LLM 案例提取、标准入库表生成和案例质量复查组织为一套可复用研究流程。适用于采集公开证据、验证机器人订单、沉淀具身智能案例库...
使用说明 (SKILL.md)

AIRS 具身智能产业研究 Skills

When to Use

当用户需要围绕具身智能、机器人产业或 embodied intelligence 做公开证据采集、订单核查、案例提取和标准化入库时,使用本 skill。

典型请求:

  • "确认这些具身智能企业的天眼查主体"
  • "采集宇树、乐聚、智元机器人的中标公告"
  • "核查第三方 Excel 里的机器人订单是否有公开招投标证据"
  • "把天眼查公告提取成具身智能案例库入库表"
  • "复查案例详情、场景分类和案例简介质量"

Capability Map

本仓库是一个多模块 skill bundle,根入口负责总调度,具体能力位于 skills/

能力 子 Skill 命令
企业主体确认 skills/company-identity/SKILL.md npm run search
中标公告采集 skills/bidding-crawl/SKILL.md npm run crawl
第三方订单核查 skills/thirdparty-verify/SKILL.md npm run verify
招投标案例提取 skills/case-extract/SKILL.md npm run extract
标准入库表生成 skills/case-ingest/SKILL.md npm run ingest
案例质量复查 skills/case-quality-review/SKILL.md npm run quality:review

如果用户只需要某个环节,优先读取对应子 Skill;如果用户要跑完整链路,按企业确认、公告采集、案例提取、人工复核、入库复查的顺序执行。

Requirements

  • Node.js 18+
  • npm
  • Chrome 浏览器
  • 天眼查账号和已登录的浏览器会话
  • LLM provider:在本地 config/settings.json 中配置 OpenAI-compatible provider,例如 Moonshot

涉及天眼查搜索、采集和核查的命令需要 Chrome 以远程调试模式启动:

/Applications/Google\ Chrome.app/Contents/MacOS/Google\ Chrome --remote-debugging-port=9222

启动后访问 https://www.tianyancha.com 并完成登录。

Setup

npm install
cp config/settings.example.json config/settings.json

在本地 config/settings.json 中填写 API key。真实 API key、天眼查登录状态、抓取结果和本地数据不要提交到仓库。

Workflow

  1. 准备企业名单模板:编辑 具身智能中游企业数据库.md
  2. 企业主体确认:运行 npm run search,生成 data/company_list.csv
  3. 公告证据采集:运行 npm run crawl,生成 data/bidding_records.csvdata/raw_content/*.md
  4. 第三方订单核查:如有外部 Excel,运行 npm run verify -- path/to/orders.xlsx
  5. 案例提取:运行 npm run extract,生成 data/review_sheet.csvdata/output/ingestion_output.csv
  6. 人工复核:在 review_sheet.csv 中确认 待验证 行,将可入库记录改为 通过,再运行 npm run extract 刷新入库输入。
  7. 标准入库和质量复查:运行 npm run ingestnpm run quality:review

Outputs

主要输出位于 data/,默认不进入发布包:

  • data/company_list.csv
  • data/bidding_records.csv
  • data/raw_content/*.md
  • data/verify_match_report.csv
  • data/verify_bidding_records.csv
  • data/extract_results.csv
  • data/review_sheet.csv
  • data/output/ingestion_output.csv
  • data/output/ingest_final.csv
  • data/output/ingest_report.md

Data And Safety Notes

  • 只发布示例配置和模板,不发布真实 Excel、抓取结果、日志、API key 或天眼查会话信息。
  • Chrome 远程调试端口只应在可信本机环境中使用;完成采集后关闭该浏览器会话。
  • 天眼查验证码或安全验证需要人工完成,不应尝试绕过平台风控。
  • 发布到 ClawHub 前检查 .clawhubignore.gitignoregit status
安全使用建议
Before installing, be comfortable running npm dependencies, creating a local config/settings.json with an API key, and starting Chrome with remote debugging. Prefer a dedicated Chrome profile for Tianyancha, close the debug browser after use, and avoid sending confidential Excel or raw-content files to an external LLM provider.
功能分析
Type: OpenClaw Skill Name: airs-embodied-intelligence-research-skills Version: 1.0.1 The skill bundle is a comprehensive research tool designed for the robotics industry to collect and analyze bidding data from the Tianyancha platform. It utilizes Puppeteer for browser automation, standard libraries for CSV/Excel processing (xlsx, csv-writer), and LLMs for structured data extraction. The code logic across files like `src/crawl_bidding.js`, `src/extract_cases.js`, and `src/utils/llm.js` is transparent, well-commented, and strictly follows the documented research workflow. There are no indicators of data exfiltration, unauthorized persistence, or malicious prompt injection; sensitive configurations like API keys are handled via a local `settings.json` file which is excluded from the repository by design.
能力标签
requires-sensitive-credentials
能力评估
Purpose & Capability
The documented capabilities—company identity checks, Tianyancha bidding crawl, order verification, LLM extraction, and CSV/report generation—fit the stated robotics industry research purpose, but they involve authenticated browsing and LLM processing.
Instruction Scope
The commands are user-directed npm scripts with documented inputs and outputs, and the docs include human review steps and CAPTCHA guidance. The scope is generally clear, but users should notice the authenticated browser automation.
Install Mechanism
There is no automatic install hook, but setup asks the user to run npm install for third-party Node dependencies. A lockfile is present and the static scan reports no suspicious patterns.
Credentials
The environment requirements are proportionate to the workflow, but Chrome remote debugging with a logged-in account and a local LLM API key are sensitive and should be used only on a trusted machine.
Persistence & Privilege
The skill writes local progress, crawl, raw-content, and output files under data/ as part of the workflow. No background service or self-persistence is shown, but the browser debug session persists until the user closes it.
如何使用
  1. 确保已安装 OpenClaw(本地或 Docker 部署)
  2. 在对话框中输入安装命令:/install airs-embodied-intelligence-research-skills
  3. 安装完成后,直接呼叫该 Skill 的名称或使用 /airs-embodied-intelligence-research-skills 触发
  4. 根据 Skill 的参数说明提供必要输入,即可获得结构化输出
版本历史
v1.0.1
Simplify published LLM provider: remove local CLI execution providers and require OpenAI-compatible API configuration in local settings.
v1.0.0
Initial release
元数据
Slug airs-embodied-intelligence-research-skills
版本 1.0.1
许可证 MIT-0
累计安装 0
当前安装数 0
历史版本数 2
常见问题

AIRS 具身智能产业研究 Skills 是什么?

AIRS 具身智能产业研究 Skills。面向 embodied intelligence、robotics 和机器人产业研究,将企业主体确认、天眼查招投标/中标公告采集、第三方订单核查、LLM 案例提取、标准入库表生成和案例质量复查组织为一套可复用研究流程。适用于采集公开证据、验证机器人订单、沉淀具身智能案例库... 它是一个面向 Claude Code / OpenClaw 的 AI Agent Skill 插件,目前累计下载 53 次。

如何安装 AIRS 具身智能产业研究 Skills?

在 OpenClaw 或 Claude Code 对话框中运行命令「/install airs-embodied-intelligence-research-skills」即可一键安装,无需额外配置。

AIRS 具身智能产业研究 Skills 是免费的吗?

是的,AIRS 具身智能产业研究 Skills 完全免费,采用 MIT-0 许可证,可自由下载、安装和使用。

AIRS 具身智能产业研究 Skills 支持哪些平台?

AIRS 具身智能产业研究 Skills 跨平台运行,可在任意部署了 OpenClaw / Claude Code 的环境中使用(cross-platform)。

谁开发了 AIRS 具身智能产业研究 Skills?

由 airs-git(@airs-guest)开发并维护,当前版本 v1.0.1。

💬 留言讨论