Deep Market Research — Development skill for Claude Code
Deep market/competitive/tech research agent skill (agentskills.io standard).
How to install Deep Market Research
This entry records only its repository, not the path inside it, so there is no
exact command to give. Open Rain3Dmetrology/deep-market-research and copy the folder into
~/.claude/skills/, or the file into ~/.claude/agents/.
What Deep Market Research does
Deep market/competitive/tech research agent skill (agentskills.io standard). Source grading + >=2-source cross-validation + dedup/staleness/fake/contradiction resolution + real-user-comment absorption. One-click install for Claude Code / OpenAI Codex / TRAE / Qodar / WorkBuddy.
Alternatives in Development
- Internet Court Skill — The trust layer for agent-to-agent commerce — natural-language mandates, ERC-7710 delegated permissions, x402 4.9k ★
- Contradiction Analysis 3.7k ★
- 08 Open Question Resolution — Open Question Resolution 1.1k ★
README
深度市场调研 Skill(Deep Market Research)
🌐 语言 / Language:**[🇨🇳 中文](README.md)** · [🇺🇸 English](README_EN.md)
跨平台 AI Agent 调研工作流:源分级 + ≥2 源交叉验证 + 去重/去旧/去假/去矛盾 + 吸收真实用户热评,输出质量稳定、可复现、带置信度标签的调研报告。
[](LICENSE)
遵循 [Agent Skills 开放标准](https://agentskills.io/)(Anthropic 发起,Claude Code / OpenAI Codex / TRAE / Qoder / WorkBuddy 等 50+ 平台原生支持)。
✨ 特性
与通用 AI 搜索 / 深度研究 skill 的核心差异:**dmr 不是搜索包装,而是一条可复现、带置信标签、终稿对抗审计的调研流水线。**
- 完整版本变更明细见 CHANGELOG.md
独有优势
- 确定性流水线:固定 Step 0–8,每次可复现、可对比
- plan-first 验收先行:Step 0 即产出验收标准(每条含度量锚点)+ 终止条件三态(达标即停 / 预算耗尽即停 / 最多 N 轮),终稿逐条对照
- 源分级置信:T1 官方 / T2 专家 / T3 二手 / T4 社媒,每条结论带置信标签
- ≥2 源交叉验证:事实拆解,冲突显式标注,不强行共识;竞品关键参数须 ≥3 源
- 终稿对抗审计:终稿前独立 critic 挑战,局部修补,不整篇重写
- 程序化机检门禁:
validate_report.py(6 条 lint)+validate_param_card.py(参数卡校验),零依赖、离线安全,交付前可机检 - 中文/CJK 原生支持:公众号、知乎、小红书、CNKI 等中文源不丢弃、不当 junk
- 零安装 Skill:纯方法论,调用 Agent 内置工具,无需额外 Python 依赖
- 可选工具永不阻断:Exa / Firecrawl / Tavily / Perplexity / GPT Researcher / ModelScope 有则增强,缺失优雅降级
- 平台无关:不绑定任何 MCP 配置 / agent-team 协议 / 专有后端,WorkBuddy / Claude / Codex / Trae / qoder / Cursor 通用
输出能力
- 五套模板:通用 / 行业赛道(麦肯锡风,含销售与商业化分析)/ 公司竞品(SWOT + 情景推演)/ 学术 / 监测增量
- intel-brief 风格:事实 → 影响 → 原因三元组织
- 学术模块:arXiv / PubMed / OpenAlex / Semantic Scholar / CNKI,优先免费 API
- 分析透镜:波特五力 / PESTEL / BCG / 3C / TAM-SOM + 销售透镜,按意图触发,不堆砌
- 研究参数卡:结构化 schema 单一权威(必含来源池 + 推荐验收标准),跨阶段整卡传递
- 增量沉淀:结构化 markdown note(YAML frontmatter)+ 实体级证据缓存,对接 ima / Obsidian / 本地 wiki
- 季度基准自测:5 题固定基准 + Q6 销售路由回归题,分数演化防质量回归
技术栈与流水线(可视化)
**调研流水线** — 主管线 Step 0–8 与三-B 深度研究闭环正交,质量由方法论保证而非某个搜索 API:

**技术栈** — 默认层零依赖零安装;可选增强层缺失即优雅降级,仅丰富素材来源:

🌐 支持的平台
| 平台 | Skills 目录 | 触发方式 |
|---|---|---|
| Claude Code / Claude | ~/.claude/skills/ |
自动发现 + /deep-market-research |
| OpenAI Codex | `~/.co |
Related Skills
Askit Build Skill
Create or improve an agentskills.io skill to the Advanced Skill Library Standard. Use to scaffold a new SKILL.
Ship CI Review Loop
Use when monitoring CI and handling review comments during /ship. Details mandatory wait periods, auto-reviewe
Aide
Pense-bête — comment utiliser le conseiller carrière tech
Stallguard
Drop-in wrapper that streams agent steps and kills/flags stalls, loops, and fake-GOAL finishes.
Discovery Prototype
Build a clickable prototype, fake door page, or Wizard of Oz console for a specific discovery question.
Paper Review And Digest
Two Claude Code skills for reading one academic paper: /paper-review (journal-club appraisal, deterministic GR
Related Agents
Askit Skill Author
Authors and improves agentskills.io skills to the Advanced Skill Library Standard. Use when delegating skill c
Copyeditor
Sole owner of deterministic quantitative checks across all Layers. Runs the 10 tools/lint.py groups (graph·hub
Industry Researcher
Industry Researcher. Owns deep competitive analysis, market sizing, trend monitoring, regulatory landscape, pe