Cue Research CLI

Cue 深度研究 CLI(@cueai/cue-cli,命令 cue)使用说明。用于行业/公司/人物/政策/合规类深度调研:一句话发起多智能体研究,输出带来源的 Markdown 报告,可落盘、可续接会话追问、可指定搭子模板。涉及「深度调研」「尽调」「行研」「公司研究」「竞品对比」「政策跟踪」「人物核查」等需要多源检索与交叉验证的场景时使用。

Install

skills CLI (Claude Code, Cursor, Codex and more)
npx skills add sensedeal/cue-skills --skill cue-research-cli
Manual install
git clone --depth 1 https://github.com/sensedeal/cue-skills.git /tmp/cue-skills
mkdir -p ~/.claude/skills
cp -r /tmp/cue-skills/cli/skills/cue-research-cli ~/.claude/skills/cue-research-cli

Personal skills live in ~/.claude/skills. Use .claude/skills in a project folder to share a skill with your team.

From the SKILL.md

# Cue 调研搭子(cue CLI)

`cue`(npm 包 `@cueai/cue-cli`)是 Cue(cuecue.cn)多智能体深度研究的本地命令行桥梁。WorkBuddy 负责安装与登录态调度,本 Skill 只负责「拿到命令后怎么调」。

**核心命令:`cue research`** —— 把一句自然语言问题交给 Cue,自动拆解子任务、多源检索交叉验证,返回带来源引用的 Markdown 研究报告。

> ⚠️ 本 skill 与仓库里那个**同名的 Python 形态** `cue-research/` 是两件事:那个是
> 让 agent 直接跑 Python 脚本的形态,本目录是 **CLI 形态**(命令 `cue`,由 WorkBuddy 安装)。
> 两者名字相近但入口不同,别混用。

## 一、何时使用 / 何时不用

**应使用**:需要多源检索 + 交叉验证 + 出报告的开放性问题,例如行业研究、公司/竞品调研、政策与监管跟踪、人物背景核查、合规风险体检、投前/贷前尽调。

**不应使用**:
- 单条事实查询(某股票现价、今天天气)—— 直接用行情/搜索工具,不要为它跑一次深度研究;
- 纯本地文件处理(解析 PDF、改写文档)—— Cue 不做这个(`cue` 没有文档解析命令);
- 需要秒级返回的交互 —— 一次研究通常数分钟。

## 二、前置条件

1. CLI 由 WorkBuddy 通过 `npm install -g @cueai/cue-cli` 安装(Node ≥ 20.12,由 WorkBuddy 托管运行时);
2. 登录态:`cue auth status`(Windows 为 `cue.cmd auth status`)输出「已登录」才算可用,退出码 0=已登录 / 2=未登录;
3. 未登录时引导用户在连接器面板点「连接」:`cue auth login` 会打印 https:// 授权 URL,由 WorkBuddy 打开浏览器完成设备码授权;
4. **浏览器授权暂不可用时的退路**:若 `cue auth login` 明确提示「服务端尚未开放设备码授权」,改用用户自填 Key —— 到 https://cuecue.cn/api-key 创建 key,然后 `cue auth login --api-key <key>`,或设 `export CUE_API_KEY=sk-...` 后重跑。**看到该提示时直接走这条退路,不要反复重试 `auth login`**;
5. 排障先看 `cue doctor`(输出 JSON:node 版本、配置文件、脱敏 key、API 连通性)。

> **登录态是怎么维持的**:设备码换来的访问令牌 1 小时过期,但 CLI 会在任何 API 调用
> 遇到 401 时**自动静默刷新一次并重试**。所以**不要因为一次 401 就急着让用户重新登录** ——
> 只有刷新也失败时,CLI 才会以退出码 2 结束,那时再引导重连。

## 三、命令速查

| 命令 | 用途 |
|---|---|
| `cue research "<问题>"` | 发起一次研究,报告打到 stdout(前台阻塞到完成) |
| `cue research "<问题>" -o <文件>` | 报告落盘为 Markdown,并附来源清单 |
Read the full skill on GitHub

Details

Skill name
cue-research-cli
Path
cli/skills/cue-research-cli/SKILL.md
License
MIT
Repo updated
11 November 2026

FAQ

What does the Cue Research CLI skill do?

Cue 深度研究 CLI(@cueai/cue-cli,命令 cue)使用说明。用于行业/公司/人物/政策/合规类深度调研:一句话发起多智能体研究,输出带来源的 Markdown 报告,可落盘、可续接会话追问、可指定搭子模板。涉及「深度调研」「尽调」「行研」「公司研究」「竞品对比」「政策跟踪」「人物核查」等需要多源检索与交叉验证的场景时使用。

How do I install the Cue Research CLI skill?

skills CLI (Claude Code, Cursor, Codex and more): npx skills add sensedeal/cue-skills --skill cue-research-cli. Skills go in ~/.claude/skills (personal) or .claude/skills (project), and Claude loads them automatically when a task matches the description.

Is Cue Research CLI popular?

It lives in sensedeal/cue-skills, which has 8 GitHub stars and 0 forks, and is included in 1 curated skill list.

Related skills

Cue Buddy

8

sensedeal/cue-skills

Use when the user wants to author / validate / debug / test / tune / pin-as-frequent a Cue 搭子(buddy) research template for a recurring scenario (corporate-credit pre-diligence, compliance snapshot, earnings review, private-fund DD, etc.) via natural conversation. Triggers: 创建搭子 / 做一个 X 搭子 / 调试模板 / 测试我的搭子 / 提交模板 / 设为常用 / design a buddy for X / mark template as frequent. Public-data tool surface only — refuse for private-data scenarios (real AML / medical diagnosis / internal accounting).

Research & ScienceIn 1 list

Cue Capital Operations

8

sensedeal/cue-skills

用 Cue 跑「资本运作」场景的深度研究:多源公开数据交叉、结论带来源。 Run Cue deep research for the "Capital Operations" scenario. 触发 Triggers: 资本运作、公司事件与资本运作、信披与监管 / capital operations, M&A, private placement

Research & ScienceIn 1 list

Cue Credit Diligence

8

sensedeal/cue-skills

用 Cue 跑「信贷尽调」场景的深度研究:多源公开数据交叉、结论带来源。 Run Cue deep research for the "Credit Diligence" scenario. 触发 Triggers: 信贷尽调、企业尽调、信披与监管 / credit diligence, pre-lending due diligence, credit risk

Research & ScienceIn 1 list

Rpk Debug

9

redpanda-data/skills

Collect Redpanda diagnostics bundles with the `rpk debug` command group, and look up bundle flags, contents, and procedures in the Redpanda docs. Use when collecting a debug bundle for Redpanda Support or self-triage, running `rpk debug bundle` on a broker host or inside a Kubernetes pod, collecting a cluster-wide bundle through the Admin API with `rpk debug remote-bundle`, or reading a bundle to triage a problem such as an Enterprise feature that stopped working. Applies to self-managed deployments only. For client-side triage of a Redpanda Cloud cluster, see /redpanda:rpk-cluster and /redpan

Software Development

Connect Debugging

9

redpanda-data/skills

Diagnoses, validates, and monitors Redpanda Connect pipelines. Use when a Redpanda Connect pipeline fails, stalls, drops messages, or won't start; when linting a config with rpk connect lint; tuning log level or format; reading Connect metrics (Prometheus, statsd, json_api) or traces (OpenTelemetry); checking /ready and /ping health endpoints; running a dry-run connection test; or diagnosing connector failures including auth errors, TLS errors, Enterprise license errors, backpressure, and checkpoint/cache issues. Covers Redpanda Connect enterprise features and their config keys: enterprise con

Software Development

Rpk Redpanda

9

redpanda-data/skills

Operate a self-managed Redpanda broker node from the CLI with the `rpk redpanda` command group and `rpk iotune`, and look up node commands and broker properties in the Redpanda docs. Use when starting, stopping, or restarting a broker, preparing a node for production, switching between production, development, and recovery mode, running the autotuner or benchmarking disks with iotune, bootstrapping or editing redpanda.yaml, or taking a broker down for a reboot or upgrade. Self-managed only: these commands act on the broker host and don't apply to Redpanda Cloud. For cluster-wide operations (cl

Software Development

Browse the most popular Claude skills, all software development skills, or write your own with the skills generator.