Borrowing it
Nothing to install: this file belongs to MaesHughes/wuxing-search-mcp. Take a copy, put it at the same path in your own repository, and replace the rules that are about this project with yours.
curl -O https://raw.githubusercontent.com/MaesHughes/wuxing-search-mcp/master/.claude/commands/search-deep.mdgit clone --depth 1 https://github.com/MaesHughes/wuxing-search-mcpWrote this? Show the measurements
A badge with what this costs and how it scanned, read live from this page, so it follows the numbers instead of freezing them. Markdown for a README, HTML for a documentation site or a project page.
[](https://agentmods.dev/commands/maeshughes/wuxing-search-mcp/search-deep)<a href="https://agentmods.dev/commands/maeshughes/wuxing-search-mcp/search-deep"><img src="https://agentmods.dev/badge/commands/maeshughes/wuxing-search-mcp/search-deep/github.svg" alt="Measured on agentmods" height="20"></a>Or the 80×15 button, for a site that already has a row of RSS and ATOM ones. Only the verdict fits; the numbers stay here.
<a href="https://agentmods.dev/commands/maeshughes/wuxing-search-mcp/search-deep"><img src="https://agentmods.dev/badge/commands/maeshughes/wuxing-search-mcp/search-deep.svg" alt="Reviewed on agentmods" width="80" height="20"></a>What it costs to keep this loaded
Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.
| Model | Per session | Once invoked |
|---|---|---|
| Fable 5.1 | $0.00000 | $0.03806 |
| Opus 5 | $0.00000 | $0.01903 |
| Sonnet 5 | $0.00000 | $0.00761 |
| Haiku 4.5 | $0.00000 | $0.00381 |
Grade A, and why
search-deep scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 11d ago.
A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.
Nothing flagged
None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.
How it starts
The opening of the file, as written. The whole thing — 332 lines — stays where its author put it; the contents beside it link to each section on GitHub.
深度搜索
你是一位 Researcher Agent(研究代理)。你的核心能力不是写文章,而是通过多轮迭代搜索,把一个问题的信息覆盖度拉到最高,然后基于充分的信息写出有深度的报告。
搜索问题
$ARGUMENTS
第一步:评估研究深度
判断问题的深度级别,不同深度决定覆盖标准(不是轮数上限):
| 深度级别 | 适用问题 | 覆盖标准 |
|---|---|---|
| 浅层 | 事实查询、简单定义 | 1-2 个维度有信息即可停止 |
| 中层 | 概念解释、对比分析、方法论 | 核心维度全部覆盖,每维度 1+ 来源 |
| 深层 | 技术选型、趋势调研、综合调研 | 7 维度全覆盖,每维度 2+ 来源 |
判断依据:
- 问题涉及几个维度?(单维度=浅层,2-3 个=中层,4+ 个=深层)
- 需要对比多个方案吗?(是=中层起步)
- 需要最新信息吗?(是=至少中层)
- 用户可能用这个信息做决策吗?(是=至少中层)
第二步:问题拆解
在开始搜索之前,必须先拆解问题,提取隐含约束。
用户的问题往往包含字面没有说出来的要求。如果你不理解这些隐含约束,搜出来的内容会跑偏。
必须输出以下拆解:
📋 问题拆解:
- 核心主题:[一句话概括用户真正要研究什么]
- 隐含约束:[从问题中推断出的隐性要求]
- 必须包含的关键词:[搜索查询中应该出现的词]
- 排除方向:[搜索中应该避免的方向/内容]
拆解示例:
用户问题:"本地知识库商业化方案调研"
📋 问题拆解:
- 核心主题:面向企业的本地部署知识库产品方案
- 隐含约束:必须本地模型(数据不出内网)、面向企业客户(不是个人玩具)、需要可交付的商业化能力
- 必须包含的关键词:本地模型、本地部署、私有化、企业级、离线、Ollama、vLLM
- 排除方向:依赖云端API(OpenAI/通义千问等)的方案、SaaS模式、个人笔记工具
用户问题:"2026年最值得学习的编程语言"
📋 问题拆解:
- 核心主题:面向求职和职业发展的编程语言选择
- 隐含约束:关注就业市场、学习成本要合理、有上升趋势
- 必须包含的关键词:编程语言、就业市场、薪资、趋势、2026
- 排除方向:已过时的语言(COBOL等)、纯学术语言(Haskell等)
拆解要点:
- 不要只看字面意思,要理解用户为什么问这个问题
- 关注修饰词:"本地""商业化""免费""开源""企业级"等都是强约束
- 关注否定意图:如果用户说"A对比B",说明用户可能倾向其中一个
- 如果问题模糊,拆解后向用户确认约束是否正确
拆解结果贯穿全程:
- 后续每一轮的搜索查询都必须符合这些约束
- 覆盖度评估时,检查搜到的内容是否满足这些约束
- 撰写报告时,所有方案都要在约束框架内评价
第三步:迭代搜索(Agent 循环)
核心机制:覆盖度驱动,不设固定轮数
没有"最多搜几轮"的限制。 每个周期(Round)执行 4 步,由覆盖度评估决定是否继续:
🔍 推理 → 🔎 搜索 → 💾 保存 → 📊 评估 → (继续 / 停止)
Step 1:强制推理
每次调用搜索工具之前,必须先输出推理过程。不要在脑中完成推理然后跳过。
格式:
🔍 推理:Round N — [当前已有什么] → [还缺什么] → [本轮搜什么]
示例:
🔍 推理:Round 3 — 已覆盖 RAGFlow/Dify 的功能对比和部署方式,但缺少:①实际用户评价 ②与商业产品的对比 ③性能基准数据。本轮搜索"RAGFlow vs Dify 用户真实评价 对比体验"。
推理输出的作用:
- 帮你理清思路,避免盲目搜索
- 让用户看到你的搜索策略,可以随时纠正方向
约束检查:每次推理时,回顾第二步的问题拆解,确认本轮查询是否满足所有约束。如果发现搜索方向偏离了约束(比如用户要求"本地模型"但搜出来的内容涉及云端 API),下一轮必须纠正方向。
Step 2:执行搜索
调用 mcp__wuxing-search__wuxing_search,传入查询。
每个 Round 聚焦一个缺口,可以搜 1-2 个相关查询来填补这个缺口。不要试图在一个 Round 里同时填补 3 个不同方向的缺口——宁可多一个 Round,每次都精准。
Step 3:保存 Research Brief
What this file has done since we first saw it
Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.
- 11d ago First seen · 332 lines · 0 tokens per session scan A 690dfa6a9c20
search-deep is a command published in the GitHub repository MaesHughes/wuxing-search-mcp (12 stars, last pushed 3mo ago), licensed MIT. It costs nothing until one of its globs matches a file; then it loads 3,806 tokens. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-30.
Other commands, from other repositories
checklist
Generate a custom checklist for the current feature based on user requirements.
clarify
Identify underspecified areas in the current feature spec by asking up to 5 highly targeted clarification questions and encoding answers back into the spec.
specify
Create or update the feature specification from a natural language feature description.
analyze
Perform a non-destructive cross-artifact consistency and quality analysis across spec.md, plan.md, and tasks.md after task generation.
converge
Assess the current codebase against the feature's spec, plan, and tasks, then append any remaining unbuilt work as new tasks to tasks.md so implement can complete it.
implement
Execute the implementation plan by processing and executing all tasks defined in tasks.md.