AIRS 具身智能产业研究 Skills

ReviewAudited by ClawScan on May 12, 2026.

Overview

This appears to be a coherent research automation skill, but it needs a logged-in Tianyancha Chrome session and an LLM API key, so run it only in a trusted local setup.

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.

Findings (4)

Artifact-based informational review of SKILL.md, metadata, install specs, static scan signals, and capability signals. ClawScan does not execute the skill or run runtime probes.

Medium

#ASI03: Identity and Privilege Abuse

What this means

The skill can use the logged-in Tianyancha session and may consume LLM API quota while processing research data.

Why it was flagged

The skill explicitly requires an authenticated Tianyancha browser session and an LLM provider API key. This is disclosed and purpose-aligned, but it grants access under the user's accounts.

Skill content

- 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

Recommendation

Use a dedicated browser profile or account where possible, keep config/settings.json private, and close the remote-debugging browser session after use.

Low

#ASI02: Tool Misuse and Exploitation

What this means

While the debug port is open, the skill can automate the logged-in browser; other local processes may also be able to access that port.

Why it was flagged

The code attaches to Chrome through the local DevTools remote-debugging port. This is expected for the Tianyancha crawling workflow, but it is a powerful browser automation channel.

Skill content

puppeteer.connect({
      browserURL: 'http://127.0.0.1:9222',
      defaultViewport: null,
    });

Recommendation

Run Chrome remote debugging only on a trusted local machine, avoid leaving the port open, and close that Chrome session when crawling is complete.

Low

#ASI07: Insecure Inter-Agent Communication

What this means

Raw announcement text or user-provided research data may be processed by the configured LLM provider.

Why it was flagged

The extraction workflow sends local crawl/raw-content data to a configured LLM provider for structured extraction. This is central to the stated purpose, but it is an external data flow if the provider is remote.

Skill content

用于读取 data/bidding_records.csv 与 data/raw_content/*.md，调用 LLM 提取准入结果、场景分类、需求方、机器人型号、金额、部署时间和案例详情

Recommendation

Do not include confidential Excel files or private raw content unless your chosen LLM provider is approved for that data.

Low

#ASI04: Agentic Supply Chain Vulnerabilities

What this means

Installing the skill fetches third-party Node packages, which carries ordinary dependency supply-chain risk.

Why it was flagged

The setup relies on npm packages for browser automation, spreadsheet parsing, CSV handling, and logging. This is normal for the skill, and a package-lock is included.

Skill content

"dependencies": {
    "puppeteer-core": "^24.0.0",
    "xlsx": "^0.18.5",
    "csv-writer": "^1.6.0",
    "csv-parse": "^5.6.0",
    "winston": "^3.17.0"
  }

Recommendation

Install from a trusted registry, keep the lockfile, and update dependencies through normal review.