← 返回 Skills 市场
2685
总下载
2
收藏
3
当前安装
13
版本数
在 OpenClaw 中安装
/install cuecue-deep-research
功能描述
专业金融深度研究工具。当用户需要了解市场行情、行业趋势、公司基本面、政策影响、竞品动态、地缘政治风险或任何需要数据支撑的金融分析时，应主动调用此技能。输出结构化、数据驱动的专业研究报告，适用于投资决策、战略规划和市场洞察等场景。
使用说明 (SKILL.md)
CueCue 深度研究技能
CueCue 是专为金融行业设计的深度研究工具。本技能说明 cue 命令行工具的调用方式，智能体应通过此技能执行金融深度研究任务。
适用场景
凡涉及以下金融研究需求，均应主动调用本技能：
- 市场调研：股票市场分析、行业趋势研判、市场预测与投资机会识别
- 行业分析：行业格局梳理、竞争动态追踪、市场结构与增长预测
- 公司研究：企业基本面分析、财务表现评估、商业模式与战略定位研究
- 政策影响评估：监管政策解读、政府措施分析及其对市场的影响评估
- 地缘政治分析：国际关系、贸易政策、区域冲突及其经济影响研判
- 竞品情报：竞争对手分析、市场定位对比、产品差异化与战略基准研究
- 舆情分析：公众舆论监测、媒体报道分析与利益相关方认知研究
- 区域市场研究：地理市场分析、区域经济状况与投资机会评估
本工具输出结构化、数据驱动的专业研究报告，为智能体和金融专业人士的决策提供可靠依据。
调用说明
本技能通过 cue 命令行工具执行研究任务。智能体应使用 exec 直接调用 cue 命令，不得使用 sessions_spawn 方式调用。
命令行参考
CLI 命令为 cue，使用 cue research \x3Cquery>（或 cue r \x3Cquery>）执行研究查询。
全局选项
| 选项 | 说明 |
|---|---|
--api-key KEY |
CueCue API 密钥（默认读取 CUECUE_API_KEY 环境变量） |
--base-url URL |
CueCue API 基础地址（默认为 https://cuecue.cn） |
cue research \x3Cquery> 选项
| 选项 | 必填 | 说明 |
|---|---|---|
query |
Y | 研究问题或主题 |
--conversation-id ID |
N | 继续已有对话 |
--template-id ID |
N | 使用预定义研究模板（不可与 --mimic-url 同时使用） |
--mimic-url URL |
N | 模仿指定 URL 的写作风格（不可与 --template-id 同时使用） |
--output, -o FILE |
Y | 将报告保存为文件（Markdown 格式）。推荐格式：~/cue-reports/YYYY-MM-DD-HH-MM-描述性名称.md（例如 ~/cue-reports/2026-01-30-12-41-tesla-analysis.md），~ 会自动展开为主目录。 |
--verbose, -v |
N | 启用详细日志 |
--foreground |
N | 在前台运行（默认：后台运行） |
--openclaw-channel CHANNEL |
Y | OpenClaw 通知渠道名称（如 feishu） |
--openclaw-channel-id ID |
Y | OpenClaw 通知渠道 ID，必须使用当前对话的 channel-id |
--help, -h |
N | 显示帮助信息 |
cue config \x3Csubcommand> 选项
| 子命令 | 说明 |
|---|---|
cue config set \x3Ckey> \x3Cvalue> |
保存配置项（如 cue config set api_key YOUR_KEY） |
cue config get [key] |
获取配置项，不指定 key 则列出所有配置 |
故障排查
401 未授权
- 确认 API 密钥填写正确
- 检查 API 密钥是否已过期
- 确认账号具有相应访问权限
连接超时
- 确认 base URL 配置正确
- 检查本地网络连接状态
- 超时不代表任务失败，研究任务可能仍在服务端处理中，可通过网页界面查看进度
报告内容为空
- 确保研究问题表述清晰、具体
- 查看服务端日志中的错误信息
- 尝试简化查询以排查连接问题
支持
如有问题或疑问：
安全使用建议
This skill appears to be what it says: a Node-based CLI wrapper that sends research queries to CueCue and writes Markdown reports to your home folder. Before installing: 1) Verify the npm package @sensedealai/cuecue and its source code or package page to ensure it matches the published project and there are no surprises. 2) Understand that using the skill will send your queries, the current conversation's channel-id/name, and any provided URLs to the external CueCue service — do not pass sensitive or private information you don't want shared. 3) Keep your CUECUE_API_KEY secret and consider using a scoped/limited key if possible. 4) Note the CLI saves files under your home directory — check the paths it will write to and clean up if needed. 5) If you need stronger guarantees, run the CLI in an isolated environment (container/sandbox) and review network traffic or the package source before granting the API key.
功能分析
Type: OpenClaw Skill
Name: cuecue-deep-research
Version: 1.1.3
The skill is a legitimate financial research tool designed to interface with the CueCue AI platform (cuecue.cn). It installs a standard npm package (@sensedealai/cuecue) and provides structured instructions for an AI agent to perform market analysis and generate reports. While it requires an API key and network access, these are consistent with its stated purpose, and no evidence of data exfiltration, unauthorized execution, or malicious prompt injection was found in SKILL.md or the associated metadata.
能力评估
Purpose & Capability
Name/description, required binary (node), primary credential (CUECUE_API_KEY), homepage (cuecue.cn), and the declared npm package (@sensedealai/cuecue) all align with a Node CLI that performs financial research. There are no unrelated credentials or unexpected system paths declared.
Instruction Scope
Runtime instructions require the agent to exec the 'cue' CLI to run research queries and save Markdown reports to the user's home folder. The skill explicitly requires passing --openclaw-channel and --openclaw-channel-id (the current dialog's channel-id) to the service and disallows sessions_spawn. These behaviors are plausible for a notification/agent-integrated research service but do mean conversation metadata and query text will be sent to the external CueCue service. The SKILL.md does not instruct reading arbitrary local files or other unrelated secrets.
Install Mechanism
Installation is via an npm package (@sensedealai/cuecue) which is a reasonable choice for a Node CLI. No arbitrary URL downloads or archive extraction are used. Two identical install entries (npm-global and npm-local) are present — redundant but not inherently malicious. npm install is moderate risk compared with no-installation skills (standard for CLIs).
Credentials
The only required environment variable is CUECUE_API_KEY, which is appropriate. However, the skill requires passing OpenClaw channel name and the current dialog's channel-id to the external service (via CLI flags), and supports a --mimic-url option which may cause the service to fetch or analyze external URLs; both can expose conversation metadata, channel identifiers, and query content to the third-party service. Confirm you trust the service with that context and protect the API key.
Persistence & Privilege
always is false and the skill does not request elevated or global persistence. Installation creates a CLI binary (cue) which is expected behavior for an npm-installed tool. The skill does not modify other skills or system-wide agent settings in the instructions.
如何使用
- 确保已安装 OpenClaw（本地或 Docker 部署）
- 在对话框中输入安装命令：
/install cuecue-deep-research - 安装完成后，直接呼叫该 Skill 的名称或使用
/cuecue-deep-research触发 - 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
v1.1.3
- 优化了技能描述和适用场景说明，使用途和调用场景更明确
- 明确要求 `cue research` 命令的 `--output`、`--openclaw-channel`、`--openclaw-channel-id` 为必填参数，并强调 channel-id 必须与当前对话一致
- 细化了命令行调用规范，强调只能用 exec 方式调用 cue 工具
- 精简并优化了故障排查和支持说明，帮助定位常见问题
- 更新了整体文档结构和表述，使内容更清晰易用
v1.1.2
- 文档补充说明了如何调用 cue research 命令，强调应使用 exec 直接调用 cue 命令行工具，而不是 sessions_spawn。
- 版本号更新至 1.1.2。
- 其它功能、命令行选项和支持内容未变。
v1.1.1
**CueCue Deep Research v1.1.1 changelog**
- Added a new README.md file for user guidance.
- SKILL.md: Major rewrite with detailed CLI options, troubleshooting, and use cases, now primarily in Chinese.
- Enhanced metadata and example usage for better integration with OpenClaw.
- Added explicit support for OpenClaw notification channels and CLI foreground/background mode.
- Updated description to emphasize broad financial research use (market intelligence, industry, company, policy, geopolitics, etc.).
v1.1.0
- Command usage guidance updated: now explicitly recommend using exec to call cue CLI directly instead of sessions_spawn.
- Improved instructions for handling 401 errors: added prompt for user registration and CUECUE_API_KEY input.
- Version updated to 1.1.0.
- Documentation clarified for CLI invocation behavior and typical execution time.
v1.0.9
- CLI 命令从 `cuecue-research` 更新为 `cue`
- SKILL.md 文档切换为简明中文说明，移除英文内容
- 增加全局与 research、config 子命令的 CLI 详细参数说明
- 新增 `--foreground`、`--openclaw-channel` 和 `--openclaw-channel-id` 选项支持
- 提供常见问题（401 未授权、超时、报告为空）和支持联系信息
- keywords、安装指令中的 CLI 名称同步为 `cue`
v1.0.8
Version 1.0.8
- Documentation updated: Added new "Advanced Options" and "Type Definitions" sections to SKILL.md with SDK usage examples.
- Expanded command-line options and output format documentation for greater clarity and technical guidance.
v1.0.7
- Updated skill version to 1.0.7.
- No functional or feature changes; documentation (SKILL.md) updated only.
v1.0.6
**Expanded description and use cases for financial research skill.**
- Improved and extended skill description for clarity and professionalism in both English and Chinese.
- Added detailed sections on when and how to use the skill, with industry-specific scenarios.
- Clarified the tool’s value for market, company, policy, competitive, sentiment, and geopolitical analysis.
- No changes to actual code or technical metadata.
- Documentation now serves both AI agents and financial professionals as users.
v1.0.5
cuecue-deep-research v1.0.5
- Added new `README.md` providing user guidance and documentation.
- Enhanced metadata in `SKILL.md` with homepage, user-invocable flag, supported OS, install instructions, and more detailed requirements.
- Updated skill description and keywords for better discoverability.
- No changes to core research functionality; improvements focus on usability and documentation.
v1.0.4
No changes detected in this version.
- Version number remains at 1.0.3.
- No file or documentation updates have been made.
- Functionality and usage remain the same.
v1.0.3
cuecue-deep-research v1.0.3
- Added detailed instructions for integrating OpenClaw's cron system for automated research progress monitoring.
- Provided clear communication guidelines emphasizing user-facing updates over technical details.
- Included example workflows and troubleshooting steps for cron-based monitoring.
- No changes to code or CLI functionality; all updates are documentation and usage guidance.
v1.0.1
**Major update: Migration to TypeScript; CLI & codebase overhaul**
- Replaced Python implementation with a TypeScript-based CLI and codebase for improved type safety and async/await usage.
- Updated documentation and usage instructions to reference Node.js ecosystem (npm, yarn) instead of Python.
- Added detailed guidelines and instructions for AI assistants and progress monitoring.
- Streamlined installation, usage, and output format descriptions for the new CLI.
- Removed Python source, sample usage, and quickstart files.
v1.0.0
cuecue-deep-research 1.0.0 – Initial Release
- Launches a command-line and Python skill to execute deep financial research with CueCue’s AI-powered multi-agent system.
- Streams only key output: task titles, the final markdown report, and a shareable report URL.
- Supports advanced options: research templates, writing style mimicry from URLs, multi-turn conversations, and report file export.
- Includes detailed CLI options, environment variable configuration, and troubleshooting guidance.
- Python API allows programmatic research workflows with async support.
元数据
常见问题
CueCue Deep Research 是什么？
专业金融深度研究工具。当用户需要了解市场行情、行业趋势、公司基本面、政策影响、竞品动态、地缘政治风险或任何需要数据支撑的金融分析时，应主动调用此技能。输出结构化、数据驱动的专业研究报告，适用于投资决策、战略规划和市场洞察等场景。 它是一个面向 Claude Code / OpenClaw 的 AI Agent Skill 插件，目前累计下载 2685 次。
如何安装 CueCue Deep Research？
在 OpenClaw 或 Claude Code 对话框中运行命令「/install cuecue-deep-research」即可一键安装，无需额外配置。
CueCue Deep Research 是免费的吗？
是的，CueCue Deep Research 完全免费，采用 MIT-0 许可证，可自由下载、安装和使用。
CueCue Deep Research 支持哪些平台？
CueCue Deep Research 跨平台运行，可在任意部署了 OpenClaw / Claude Code 的环境中使用（darwin, linux）。
谁开发了 CueCue Deep Research？
由 Chris（@xfgong）开发并维护，当前版本 v1.1.3。
推荐 Skills