Fxbaogao Deep Research

围绕行业、公司或主题做深度研究。采用"先想后搜"方法：先明确研究边界，用 MECE 原则拆解关键子问题，再基于发现报告接口定向搜索、筛选核心研报、精读后沉淀事实、对比推导，最终形成结构化结论。

Install

skills CLI (Claude Code, Cursor, Codex and more)

npx skills add fxbaogao/fxbaogao-deep-research --skill fxbaogao-deep-research

Manual install

git clone --depth 1 https://github.com/fxbaogao/fxbaogao-deep-research.git /tmp/fxbaogao-deep-research
mkdir -p ~/.claude/skills
cp -r /tmp/fxbaogao-deep-research/. ~/.claude/skills/fxbaogao-deep-research

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

From the SKILL.md

# 发现报告深度研究 Skill

## 何时使用

- 用户要研究行业、公司或主题，并需要系统收集研报与结构化分析。
- 用户希望从发现报告中筛选核心观点、关键事实、数据口径和分歧。
- 用户明确要求基于发现报告、fxbaogao、研报、报告 ID 或 PDF 做研究。

## 先读什么

- 调用任何发现报告接口前，先读 `references/api-usage.md`。
- 做多篇报告的完整深度研究时，读 `references/research-playbook.md`。
- 做单篇报告精读时，读 `references/report-reader-playbook.md`。
- 需要落盘时，复用 `assets/workspace_templates/` 下的模板结构。

## 授权与凭据

网页授权、API Key 的本地存储位置和安全约束统一以 `references/api-usage.md` 的“网页授权”为准。调用任何需要鉴权的接口前，必须按该流程检查或获取凭据；不得要求用户复制、发送或手动配置 API Key。

## 路由规则

### 单篇报告精读

当用户给出明确报告 ID、单篇 PDF、单篇报告标题，或明确要求“精读某一篇报告”时，直接进入单篇精读流程。

- 不默认创建完整多篇研究工作区。
- 不要求先做 MECE 子问题搜索矩阵。
- 不额外搜索候选研报，除非用户要求补充验证。
- 输出围绕该报告：基础信息、核心逻辑、关键证据、假设、风险、事实卡片、待外部验证点、精读判断。

### 多篇深度研究

当用户提出行业、公司或主题研究时，使用完整研究流程：

1. 定义研究主题和边界。
2. 拆解 3-7 个 MECE 子问题，并为每个子问题给出搜索关键词。
3. 按子问题搜索候选研报。
4. 筛选核心必读、补充参考和背景材料。
5. 拉取核心报告段落；需要引用原文或精读时下载 PDF 核对。
6. 精读核心报告，沉淀事实卡片。
7. 建立对比框架和推导过程。
8. 输出最终调研报告。

硬性门槛：调用搜索接口前，必须先产出可审阅的研究边界和 MECE 子问题搜索矩阵。如果用户没有给出边界，先做保守默认并写明；只有边界会显著改变结论时才追问。

## 工作区

完整研究需要落盘时，创建：

```text
workspace/topic-<主题>-<日期>/
├── REPORT_READER_精读分析.md
├── FINAL_调研报告.md
└── intermediate/
    ├── 00_问题拆解.md
    ├── 01_资料来源.md
    ├── 02_事实卡片.md
    ├── 03_对比框架.md
    └── 04_推导过程.md
```

模板位于 `assets/workspace_templates/`。文件内容规范和研究细节见 `references/research-playbook.md`。

## 接口与证据约束

- 搜索、段落、下载接口以 `references/api-usage.md` 为准。
- 发现报告资料必须可追溯到报告 ID 和官网阅读链接：`https://www.fxbaogao.com/view?id=<reportId>`。
- paragraph 回包只适合快速初筛和轻量研究。深度研究、原文引用、关键数据核验必须下载 PDF 并基于 PDF 原文核对。
- 不伪造作者、机构、日期、摘要、页码或数据口径。

Read the full skill on GitHub

Details

Skill name: fxbaogao-deep-research
Repository: fxbaogao/fxbaogao-deep-research
Repo updated: 24 July 2026

FAQ

What does the Fxbaogao Deep Research skill do?

How do I install the Fxbaogao Deep Research skill?

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

Is Fxbaogao Deep Research popular?

It lives in fxbaogao/fxbaogao-deep-research, which has 10 GitHub stars and 1 forks.

Related skills

Paper Quality Filter

yilaai/paper-quality-filter

Use when screening, ranking, prioritizing, or explaining academic papers by publication venue quality. This bilingual skill helps AI agents evaluate paper lists, references, search results, BibTeX/RIS-like text, CSV, JSON, Markdown, or DOI metadata using local evidence such as JCR quartiles, impact factors, CAS zones, CCF conference ranks, EI indexing, and Chinese core journal tags including 北大核心, CSSCI, AMI核心, and 科技核心. Use also for Chinese requests such as 筛选高质量论文, 核心期刊筛选, 论文分区判断, CCF会议等级, EI收录, 北大核心, CSSCI, 科技核心, 文献综述筛文献.

Research & Science

Apple Health Analysis

labrinyang/apple-health-analysis

Clinical-grade deep analysis of Apple Health export data. Parses the XML from iPhone's Health app export and produces a comprehensive clinical-quality health assessment report using 20+ peer-reviewed statistical methods including Granger Causality, Transfer Entropy, Convergent Cross Mapping, Sample Entropy, DFA, Cosinor analysis, Kovatchev glucose risk indices, Bayesian change-point detection, and biological age estimation. Use this skill whenever the user mentions Apple Health data, health export, health XML, CGM data, wants to analyze their fitness/sleep/heart rate/glucose/weight trends, or

Research & Science

Clinical Calculator

JuneYaooo/clinical-calculator

Use for searching, selecting, calculating, checking, or explaining executable clinical formulas, medical calculators, risk and staging scores, renal estimates, lab-derived indices, unit conversions, dose arithmetic, or interpretation thresholds; also use to turn a supplied JSON, CSV, Markdown, PDF, DOCX, or written specification into a draft, validated, and installed user-defined scalar/multi-output formula, lookup table, or decision-tree calculator. Supports Chinese and English calculator names and keeps local executability separate from clinical review and release.

Research & Science

Deep Research

24601/agent-deep-research

Async deep research via Gemini Interactions API (no Gemini CLI dependency). RAG-ground queries on local files (--context), preview costs (--dry-run), structured JSON output, adaptive polling. Universal skill for 30+ AI agents including Claude Code, Amp, Codex, and Gemini CLI.

Research & Science

Paperless Ngx

ColCh/paperless-ngx-skill

Use for Paperless-NGX API work: search, upload, download, metadata, tags, status, logs, bulk ops. Triggers: paperless.

Research & Science

Math Research Activator

the-thinker0/math-skill

数学研究助手：将 AI 研究目标转成数学对象、可实现机制和可验证结论；用于架构/算子构造、理论分析、跨数学领域思路发现与结构迁移、研究相关数学查询，以及密码学定义、构造、归约与协议审查。纯实现型 debug、重构、调参和一般代码审查不触发。 English: Turn AI research goals into mathematical formulations, implementable mechanisms, and testable conclusions. Use for architecture/operator construction, theoretical analysis, cross-field mathematical discovery, math-to-AI transfer, research-related mathematics, and cryptographic definitions, constructions, reductions, or protocol reviews. Exclude implementation-only debugging, refactoring, tuning, and general code review.

Research & Science

Browse the most popular Claude skills, all research & science skills, or write your own with the skills generator.