← 返回 Skills 市场
zhangsaizz

Multi Ai Search Analysis

作者 zhangsaizz · GitHub ↗ · v1.0.2 · MIT-0
cross-platform ⚠ suspicious
190
总下载
0
收藏
1
当前安装
3
版本数
在 OpenClaw 中安装
/install multi-ai-search-analysis
功能描述
同时调用多家 AI 独立分析复杂问题,交叉验证数据,综合多视角提炼对比分析报告。
安全使用建议
What to check before installing or running: 1) Don't run this in your primary machine: run inside an isolated VM/container or sandbox. 2) Inspect scripts/login.py, scripts/run.py, scripts/extractor.py for any network endpoints or unexpected POST/GET calls and for any code that reads local filesystem paths beyond the included browser-profile. 3) Remove or replace the provided ./browser-profile before first run (the repo's profile may contain cookies/config); start with a fresh Playwright profile. 4) Grep the repo for any hardcoded remote endpoints, secrets, or telemetry URLs; verify no private upload/exfil endpoints exist. 5) If you need persistent login, log in interactively rather than importing someone else's profile. 6) Prefer running with headless=false and watch the automation the first few runs to confirm it only opens the documented AI chat pages. 7) If you are not comfortable auditing the code, avoid installing or run only the documentation/instruction parts; do not provide any credentials or copy sensitive profiles into the project's userDataDir. Additional information that would raise or lower confidence: seeing that the included browser-profile is empty/sanitized and login automation does not read autofill/storage would lower risk; finding scripts that read cookies, local storage, or send extracted credentials to external endpoints would make this clearly malicious.
功能分析
Type: OpenClaw Skill Name: multi-ai-search-analysis Version: 1.0.2 The skill bundle is a comprehensive browser automation tool designed to query multiple AI platforms (DeepSeek, Qwen, Doubao, Kimi, and Gemini) in parallel and generate comparative analysis reports. It utilizes Playwright for browser orchestration (`scripts/run.py`), includes a data extraction engine using regex (`scripts/extractor.py`), and a report generator (`scripts/reporter.py`). The bundle is well-documented with extensive installation, testing, and usage guides. While it includes a pre-configured browser profile directory (`browser-profile/`), the contents appear to be standard Chromium/Edge configuration files (such as IE compatibility lists) and do not contain evidence of malicious intent, data exfiltration, or unauthorized persistence mechanisms.
能力评估
Purpose & Capability
The name/description and SKILL.md describe browser automation to query multiple chat AIs and aggregate answers, which matches the provided scripts (run.py, reporter.py, extractor.py). That core capability is coherent. However, the repository also contains a pre-populated browser-profile directory and EntityExtraction/autofill mappings that are not explained in the docs and look unrelated to simply asking AI chat pages for responses.
Instruction Scope
SKILL.md instructs creating a persistent browser context (userDataDir ./browser-profile) and automating login checks and snapshots of responses. While prompting the user to log in is expected, the instructions + code imply reading/writing a persistent profile on disk and using selector-based extraction. The repo's browser-profile contains broad site lists and entity/autofill configs (credit card, password, account fields), which go beyond extracting AI replies and could be used to access or scrape sensitive site data if the automation ever visits other pages or if the profile already contains cookies/credentials.
Install Mechanism
There is no formal install spec (no remote downloads), which reduces supply-chain risk, but the package includes an embedded browser-profile directory that will be used as userDataDir. Shipping a pre-filled browser profile in the repo is risky because it may contain cookies, session state, or intrusive config. The code expects Playwright and a Python env (requirements.txt/install.py), so running install scripts will create/modify a local browser context on disk.
Credentials
The skill requests no environment variables or API keys, which is proportional. However, the included files reveal capabilities that could access highly sensitive data: EntityExtraction/domains_config.json (autofill mappings including passwords, credit card fields), SiteList.xml and many domain entries (banks, enterprise sites), and browser arbitration/config files. Those artifacts are not justified by the stated purpose and increase the risk that sensitive data could be read from a persistent profile or misused by automation.
Persistence & Privilege
The config explicitly sets userDataDir to ./browser-profile and the SKILL.md and scripts emphasize persistent login state ('save login state', 'persistent context'). Persistent profiles are useful for this task but increase blast radius: cookies, auth sessions, and saved autofill data in the provided browser-profile could be exposed or reused. The repo includes a pre-populated profile rather than a clean template, which is more privilege than the skill's description requires.
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 multi-ai-search-analysis
  3. 安装完成后,直接呼叫该 Skill 的名称或使用 /multi-ai-search-analysis 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
v1.0.2
- 首个版本上线。 - 支持多家 AI (DeepSeek、Qwen、豆包、Kimi、Gemini) 独立分析同一复杂问题,并汇总对比报告。 - 提供标准化问题模板与操作流程,确保分析结果一致、可比。 - 文档详细说明适用场景、最佳实践和注意事项。 - 支持串行与并行两种执行模式。
v1.0.1
v1.0.1 - No code or documentation changes detected. - This is a maintenance release with no user-facing updates.
v1.0.0
multi-ai-search-analysis v1.0.0 - 初始版本发布,支持多家 AI(DeepSeek、Qwen、豆包、Kimi、Gemini)独立分析同一复杂主题并综合生成对比报告 - 明确标准流程:统一问题表述、依次/并行询问、汇总关键信息与生成多维度对比报告 - 提供标准化操作指引,包括会话管理、平台选择、数据记录与汇总模板 - 包含常用分析场景、示例命令与推荐平台组合 - 强调统一输入、干净上下文及独立分析以确保报告客观性与可比性
元数据
Slug multi-ai-search-analysis
版本 1.0.2
许可证 MIT-0
累计安装 1
当前安装数 1
历史版本数 3
常见问题

Multi Ai Search Analysis 是什么?

同时调用多家 AI 独立分析复杂问题,交叉验证数据,综合多视角提炼对比分析报告。 它是一个面向 Claude Code / OpenClaw 的 AI Agent Skill 插件,目前累计下载 190 次。

如何安装 Multi Ai Search Analysis?

在 OpenClaw 或 Claude Code 对话框中运行命令「/install multi-ai-search-analysis」即可一键安装,无需额外配置。

Multi Ai Search Analysis 是免费的吗?

是的,Multi Ai Search Analysis 完全免费,采用 MIT-0 许可证,可自由下载、安装和使用。

Multi Ai Search Analysis 支持哪些平台?

Multi Ai Search Analysis 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 Multi Ai Search Analysis?

由 zhangsaizz(@zhangsaizz)开发并维护,当前版本 v1.0.2。

💬 留言讨论