gemini-search

Compare original and translation side by side

🇺🇸

Original

English

🇨🇳

Translation

Chinese

Gemini Literature Search

Gemini文献搜索

Search query: $ARGUMENTS

搜索查询：$ARGUMENTS

Role & Positioning

角色与定位

This skill uses Gemini as a broad literature discovery source:

Skill	Source	Best for
`/arxiv`	arXiv API	Latest preprints, cutting-edge unrefereed work
`/semantic-scholar`	Semantic Scholar API	Published venue papers (IEEE, ACM, Springer) with citation counts
`/deepxiv`	DeepXiv CLI	Layered reading: search, brief, section map, section reads
`/exa-search`	Exa API	Broad web search: blogs, docs, news, companies, research papers
`/gemini-search`	Gemini MCP / CLI	AI-powered broad literature discovery — searches across multiple angles, aliases, and sub-problems

Use Gemini when you want AI-driven discovery that goes beyond keyword matching — Gemini decomposes topics into sub-problems, explores naming variants, and surfaces papers that traditional API searches may miss.

本技能将Gemini作为广泛文献发现的来源：

技能	数据源	适用场景
`/arxiv`	arXiv API	最新预印本、前沿未评审成果
`/semantic-scholar`	Semantic Scholar API	已发表期刊/会议论文（IEEE、ACM、Springer），附带引用量
`/deepxiv`	DeepXiv CLI	分层阅读：搜索、摘要、章节地图、章节精读
`/exa-search`	Exa API	广泛网页搜索：博客、文档、新闻、企业信息、研究论文
`/gemini-search`	Gemini MCP / CLI	AI驱动的广泛文献发现 —— 从多个角度、别名和子问题进行搜索

当你需要超越关键词匹配的AI驱动发现时使用Gemini——Gemini会将主题分解为子问题，探索命名变体，挖掘传统API搜索可能遗漏的论文。

Constants

常量定义

MAX_RESULTS = 15 — Target number of papers Gemini should find.
MIN_YEAR = 2022 — Default minimum publication year. Override with
```
— year: 2020-
```
.
DEFAULT_MODEL = gemini-3-pro-preview — Strongest available Gemini option (Gemini 3 Pro). Requires
```
gemini-cli
```
v0.40+ and
```
mcp__gemini-cli__ask-gemini
```
accepting Gemini 3 aliases (verified). Override with
```
— model: gemini-3-flash-preview
```
(Gemini 3 Flash, faster, higher quota),
```
— model: auto-gemini-3
```
(auto-routes inside the Gemini 3 family by load), or
```
— model: gemini-2.5-pro
```
/
```
gemini-2.5-flash
```
(legacy, for users on older
```
gemini-cli
```
< v0.40). The MCP tool accepts all of these verbatim.

Overrides (append to arguments):
/gemini-search "topic" — max: 20
— request up to 20 papers
/gemini-search "topic" — year: 2020-
— papers from 2020 onward
/gemini-search "topic" — code-only
— only papers with open-source code
/gemini-search "topic" — venues: NeurIPS,ICML,ICLR
— focus on specific venues
/gemini-search "topic" — model: gemini-3-flash-preview
— Gemini 3 Flash (faster, higher quota, less capable than Pro)
/gemini-search "topic" — model: auto-gemini-3
— auto-routes within the Gemini 3 family by load
/gemini-search "topic" — model: gemini-2.5-pro
— legacy (only if your
gemini-cli
< v0.40)

MAX_RESULTS = 15 —— Gemini需查找的目标论文数量。
MIN_YEAR = 2022 —— 默认最低发表年份。可通过
```
— year: 2020-
```
覆盖。
DEFAULT_MODEL = gemini-3-pro-preview —— 目前最强的Gemini选项（Gemini 3 Pro）。需要
```
gemini-cli
```
v0.40+，且
```
mcp__gemini-cli__ask-gemini
```
支持Gemini 3别名（已验证）。可通过
```
— model: gemini-3-flash-preview
```
（Gemini 3 Flash，速度更快、配额更高）、
```
— model: auto-gemini-3
```
（根据负载自动在Gemini 3系列中路由）或
```
— model: gemini-2.5-pro
```
/
```
gemini-2.5-flash
```
（旧版本，适用于使用
```
gemini-cli
```
< v0.40的用户）覆盖。MCP工具会直接接受这些参数。

参数覆盖（追加到查询参数后）：
/gemini-search "主题" — max: 20
—— 请求最多20篇论文
/gemini-search "主题" — year: 2020-
—— 查找2020年及以后的论文
/gemini-search "主题" — code-only
—— 仅查找包含开源代码的论文
/gemini-search "主题" — venues: NeurIPS,ICML,ICLR
—— 聚焦特定会议/期刊
/gemini-search "主题" — model: gemini-3-flash-preview
—— 使用Gemini 3 Flash（速度更快、配额更高，能力略逊于Pro版本）
/gemini-search "主题" — model: auto-gemini-3
—— 根据负载自动在Gemini 3系列中路由
/gemini-search "主题" — model: gemini-2.5-pro
—— 旧版本（仅适用于
gemini-cli
< v0.40的用户）

Environment & Setup

环境与配置

Prerequisites

前提条件

Node.js v16.0.0+
Google Gemini CLI — installed and authenticated
bash
```
npm install -g @google/gemini-cli
gemini auth
```
gemini-mcp-tool — MCP bridge for Claude Code (jamubc/gemini-mcp-tool)
bash
```
npm install -g gemini-mcp-tool
```

Node.js v16.0.0+
Google Gemini CLI —— 已安装并完成认证
bash
```
npm install -g @google/gemini-cli
gemini auth
```
gemini-mcp-tool —— 用于Claude Code的MCP桥接工具（jamubc/gemini-mcp-tool）
bash
```
npm install -g gemini-mcp-tool
```

MCP Configuration

MCP配置

~/.claude.json

(or

%APPDATA%\Claude\claude_desktop_config.json

for Claude Desktop), add:

json

{
  "mcpServers": {
    "gemini-cli": {
      "command": "gemini-mcp"
    }
  }
}

Alternative via

npx

(auto-install):

json

{
  "mcpServers": {
    "gemini-cli": {
      "command": "npx",
      "args": ["-y", "gemini-mcp-tool"]
    }
  }
}

Or one-line setup:

bash

claude mcp add gemini-cli -- npx -y gemini-mcp-tool

在

~/.claude.json

（Claude桌面版为

%APPDATA%\Claude\claude_desktop_config.json

）中添加：

json

{
  "mcpServers": {
    "gemini-cli": {
      "command": "gemini-mcp"
    }
  }
}

通过

npx

自动安装的替代方案：

json

{
  "mcpServers": {
    "gemini-cli": {
      "command": "npx",
      "args": ["-y", "gemini-mcp-tool"]
    }
  }
}

或一键配置：

bash

claude mcp add gemini-cli -- npx -y gemini-mcp-tool

Authentication

认证

Gemini CLI uses your Google account or an API key. Add to

.claude/.env

bash

undefined

Gemini CLI使用你的Google账户或API密钥。将密钥添加到

.claude/.env

：

bash

undefined

.claude/.env

GEMINI_API_KEY=your-key-here


Claude Code automatically loads `.claude/.env` as environment variables.

- Free key from [Google AI Studio](https://aistudio.google.com/apikey)
- Flash model (`gemini-2.5-flash`) has a generous free tier (500 req/min)

GEMINI_API_KEY=your-key-here


Claude Code会自动加载`.claude/.env`作为环境变量。

- 可从[Google AI Studio](https://aistudio.google.com/apikey)获取免费密钥
- Flash模型（`gemini-2.5-flash`）提供慷慨的免费额度（500次请求/分钟）

Available MCP Tools

可用MCP工具

Tool	Parameters	Description
`mcp__gemini-cli__ask-gemini`	`prompt` (required), `model` (optional), `sandbox` (optional)	Ask Gemini for analysis or research; supports `@file` syntax
`mcp__gemini-cli__sandbox-test`	`prompt` (required), `model` (optional)	Safe code execution in sandbox
`mcp__gemini-cli__ping`	—	Connection test
`mcp__gemini-cli__help`	—	Show Gemini CLI help

工具	参数	描述
`mcp__gemini-cli__ask-gemini`	`prompt` （必填）, `model` （可选）, `sandbox` （可选）	向Gemini请求分析或研究；支持 `@file` 语法
`mcp__gemini-cli__sandbox-test`	`prompt` （必填）, `model` （可选）	在沙箱中安全执行代码
`mcp__gemini-cli__ping`	—	连接测试
`mcp__gemini-cli__help`	—	显示Gemini CLI帮助信息

Verify Setup

验证配置

bash

gemini --version

bash

gemini --version

Workflow

工作流程

Step 1: Parse Arguments

步骤1：解析参数

Parse

$ARGUMENTS

for:

query: The research topic (required)
max: Override MAX_RESULTS
year: Minimum publication year (e.g.,
```
2020-
```
)
code-only: Only include papers with open-source code
venues: Comma-separated venue filter
model: Override DEFAULT_MODEL

解析

$ARGUMENTS

中的内容：

query: 研究主题（必填）
max: 覆盖MAX_RESULTS
year: 最低发表年份（例如：
```
2020-
```
）
code-only: 仅包含带有开源代码的论文
venues: 逗号分隔的会议/期刊筛选条件
model: 覆盖DEFAULT_MODEL

Step 2: Execute Search (MCP Priority)

步骤2：执行搜索（优先使用MCP）

Priority 1 — Gemini MCP (preferred):

Try calling

mcp__gemini-cli__ask-gemini

with the search prompt:

mcp__gemini-cli__ask-gemini({
  prompt: 'You are a research literature scout. Search comprehensively for papers on: "QUERY"

IMPORTANT CONSTRAINTS:
1. Search from MULTIPLE angles — do not just use the exact query. Decompose the topic into sub-problems, aliases, neighboring tasks, and common benchmark/settings variants.
2. Prefer papers that are genuinely relevant, not merely keyword-adjacent.
3. Include top venues, journals, surveys, recent preprints, and papers with code when available.
4. Focus on papers from MIN_YEAR onward unless older foundational work is necessary.

For EACH paper found, provide ALL of the following in this exact format:
- Title: [exact title]
- Authors: [full author list]
- Year: [publication year]
- Venue: [exact conference/journal name + year, or "arXiv preprint" if not published]
- arXiv ID: [format 2401.12345, or "N/A"]
- DOI: [if available, or "N/A"]
- Code URL: [GitHub/GitLab link if available, or "No code"]
- Summary: [one-sentence core contribution]

Find at least MAX_RESULTS papers with good coverage across:
- strong recent papers from top venues
- surveys/reviews if they exist
- papers with open-source code
- closely related variants of the topic

Format as a numbered list with all fields for each paper.',
  model: 'DEFAULT_MODEL'
})

Priority 2 — Gemini CLI fallback (if MCP unavailable):

mcp__gemini-cli__ask-gemini

fails or is not configured, fall back to CLI:

bash

gemini -p 'You are a research literature scout. Search comprehensively for papers on: "QUERY"
...same prompt as above...' 2>/dev/null

Timeout: 120 seconds
Stderr: Pipe to
```
/dev/null
```
— contains hook warnings, not part of the response

When to use which:

MCP is preferred because it integrates natively with Claude Code's tool system, handles model selection, and avoids shell escaping issues.
CLI fallback ensures the skill works even when MCP is not configured or the MCP server process has crashed.

优先级1 —— Gemini MCP（推荐）：

尝试调用

mcp__gemini-cli__ask-gemini

并传入搜索提示：

mcp__gemini-cli__ask-gemini({
  prompt: '你是一名研究文献侦察员。请全面搜索关于以下主题的论文："QUERY"

重要约束：
1. 从多个角度搜索——不要仅使用精确查询。将主题分解为子问题、别名、相关任务以及常见基准/设置变体。
2. 优先选择真正相关的论文，而非仅关键词匹配的论文。
3. 涵盖顶级会议/期刊、综述、最新预印本，以及带有代码的论文（如有）。
4. 重点关注MIN_YEAR及以后的论文，除非需要引用更早的基础性成果。

对于找到的每一篇论文，请严格按照以下格式提供所有信息：
- Title: [精确标题]
- Authors: [完整作者列表]
- Year: [发表年份]
- Venue: [精确会议/期刊名称+年份，未发表则为"arXiv preprint"]
- arXiv ID: [格式为2401.12345，无则为"N/A"]
- DOI: [如有则提供，无则为"N/A"]
- Code URL: [GitHub/GitLab链接（如有），无则为"No code"]
- Summary: [一句话核心贡献]

至少找到MAX_RESULTS篇论文，确保覆盖以下类型：
- 顶级会议近期优质论文
- 相关综述（如有）
- 带有开源代码的论文
- 主题的密切相关变体

将结果格式化为编号列表，每篇论文包含所有字段。',
  model: 'DEFAULT_MODEL'
})

优先级2 —— Gemini CLI fallback（当MCP不可用时）：

如果

mcp__gemini-cli__ask-gemini

调用失败或未配置，则回退到CLI：

bash

gemini -p '你是一名研究文献侦察员。请全面搜索关于以下主题的论文："QUERY"
...上述相同提示内容...' 2>/dev/null

超时设置: 120秒
标准错误输出: 重定向到
```
/dev/null
```
——包含钩子警告，不属于响应内容

使用场景区分：

MCP是首选方案，因为它能原生集成到Claude Code的工具系统中，处理模型选择，并避免shell转义问题。
CLI回退确保即使MCP未配置或MCP服务器进程崩溃，该技能仍能正常工作。

Step 3: Parse Results

步骤3：解析结果

Extract structured paper information from Gemini's response. For each paper, normalize to:

{
  title, authors, year, venue,
  arxiv_id,    // "N/A" if not available
  doi,         // "N/A" if not available
  code_url,    // "No code" if not available
  summary      // one-sentence contribution
}

If Gemini returns fewer papers than requested, note this but do not re-query.

从Gemini的响应中提取结构化的论文信息。将每篇论文标准化为以下格式：

{
  title, authors, year, venue,
  arxiv_id,    // 无则为"N/A"
  doi,         // 无则为"N/A"
  code_url,    // 无则为"No code"
  summary      // 一句话核心贡献
}

如果Gemini返回的论文数量少于请求数量，只需注明此情况，无需重新查询。

Step 4: Present Results

步骤4：展示结果

Format results as a structured table:

| # | Title | Venue | Year | Code | Summary |
|---|-------|-------|------|------|---------|
| 1 | ... | NeurIPS 2024 | 2024 | [GitHub](url) | ... |
| 2 | ... | IEEE TWC | 2023 | No | ... |

For each paper, also show:

arXiv ID: if available (for cross-reference with
```
/arxiv
```
)
DOI: if available (canonical link for published papers)
Code: GitHub/GitLab link or "No"

将结果格式化为结构化表格：

| # | 标题 | 会议/期刊 | 年份 | 代码 | 摘要 |
|---|-------|-------|------|------|---------|
| 1 | ... | NeurIPS 2024 | 2024 | [GitHub](链接) | ... |
| 2 | ... | IEEE TWC | 2023 | 无 | ... |

对于每篇论文，还需展示：

arXiv ID: 如有（用于与
```
/arxiv
```
交叉引用）
DOI: 如有（已发表论文的标准链接）
代码: GitHub/GitLab链接或“无”

Step 5: Offer Follow-up

步骤5：提供后续操作建议

After presenting results, suggest:

text

/semantic-scholar "topic"    — search published venue papers with citation counts
/arxiv "arXiv:XXXX.XXXXX"   — fetch specific preprint details
/research-lit "topic" — sources: gemini, semantic-scholar  — combined multi-source review
/novelty-check "idea"       — verify novelty against literature

展示结果后，建议用户：

text

/semantic-scholar "主题"    —— 搜索已发表会议/期刊论文，附带引用量
/arxiv "arXiv:XXXX.XXXXX"   —— 获取特定预印本详情
/research-lit "主题" — sources: gemini, semantic-scholar  —— 多来源综合文献综述
/novelty-check "想法"       —— 验证想法的创新性

Key Rules

核心规则

MCP first, CLI second. Always try
```
mcp__gemini-cli__ask-gemini
```
before falling back to
```
gemini -p
```
.
Gemini is a discovery source, not a database. Its results may include papers it "knows about" from training data. Always cross-verify critical details (exact titles, venues, years) via
```
/semantic-scholar
```
or
```
/arxiv
```
when precision matters.
Do not use Gemini for citation counts. It may hallucinate citation numbers. Use Semantic Scholar for authoritative citation data.
Pipe stderr to
/dev/null
in CLI mode — Gemini CLI emits hook warnings on stderr.
Timeout generously in CLI mode — Gemini's thorough search can take 30-60 seconds. Set timeout to 120s.
If both MCP and CLI are unreachable, suggest using
```
/semantic-scholar
```
,
```
/arxiv
```
, or
```
/research-lit "topic" — sources: web
```
as alternatives.

优先使用MCP，其次是CLI。始终先尝试
```
mcp__gemini-cli__ask-gemini
```
，再回退到
```
gemini -p
```
。
Gemini是发现工具，而非数据库。其结果可能包含训练数据中“已知”的论文。当需要精确信息时，务必通过
```
/semantic-scholar
```
或
```
/arxiv
```
交叉验证关键细节（如精确标题、会议/期刊、年份）。
不要使用Gemini获取引用量。它可能生成虚假的引用数字。如需权威引用数据，请使用Semantic Scholar。
CLI模式下将标准错误输出重定向到
/dev/null
——Gemini CLI会在标准错误输出中发送钩子警告。
CLI模式下设置充足的超时时间——Gemini的全面搜索可能需要30-60秒。设置超时为120秒。
如果MCP和CLI均无法访问，建议用户使用
```
/semantic-scholar
```
、
```
/arxiv
```
或
```
/research-lit "主题" — sources: web
```
作为替代方案。