setup-tooluniverse by mims-harvard/tooluniverse
npx skills add https://github.com/mims-harvard/tooluniverse --skill setup-tooluniverse引导用户逐步完成 ToolUniverse 的设置。
ToolUniverse 拥有 1200 多个工具。tooluniverse 命令会自动启用精简模式,仅暴露 5 个核心 MCP 工具(list_tools, grep_tools, get_tool_info, execute_tool, find_tools),同时通过 execute_tool 保持所有工具的可访问性。
始终先用简明语言解释:
ToolUniverse 是免费的开源软件,可连接 2000 多个科学数据库(PubMed、UniProt、ChEMBL、FAERS、ClinicalTrials.gov 等)。您无需访问每个网站,只需从一个地方进行搜索。可以把它想象成科学数据库的通用遥控器。
为什么需要 AI 助手? AI 会解读您的问题,确定要搜索哪些数据库,执行查询并总结结果。您只需提问即可。
使用 AskQuestion 呈现:
| 模式 | 含义 | 适用人群 |
|---|---|---|
| 聊天模式 |
广告位招租
在这里展示您的产品或服务
触达数万 AI 开发者,精准高效
| 向 AI 助手提问。无需编码。 |
| 大多数研究人员。 |
| 命令行 | 在终端中输入简短命令。 | 快速测试。熟悉终端的用户。 |
| Python 代码 | 为自动化流程编写脚本。 | 程序员。 |
选项:"我想提问" → 聊天模式 | "快速尝试" → CLI | "我写 Python" → SDK | "我不知道" → 推荐聊天模式
如果是聊天模式,询问使用哪个应用(AskQuestion):Cursor、Claude Desktop、VS Code/Copilot、Windsurf、Claude Code、Gemini CLI、Codex、Cline/Trae/Antigravity/OpenCode。"我没有任何这些应用" → 推荐 Claude Desktop。
唯一前提条件:uv(自动管理其他所有内容)。
终端帮助(如果需要):Mac:Cmd+空格 → 输入"Terminal" → 回车。Windows:Win 键 → 输入"PowerShell" → 回车。
curl -LsSf https://astral.sh/uv/install.sh | sh
(这是一个安全、标准的命令,用于下载并安装 uv,一个轻量级包管理器。它被 Python 开发者广泛使用。安装完成后,请关闭并重新打开终端。)
验证:uv --version
确保步骤 2 已完成,然后尝试:
uvx --from tooluniverse tu status # 有多少工具?
uvx --from tooluniverse tu find 'drug safety' # 按主题搜索
uvx --from tooluniverse tu info FAERS_count_death_related_by_drug # 查看参数
uvx --from tooluniverse tu run FAERS_count_death_related_by_drug '{"drug_name": "metformin"}'
首次运行约需 30 秒(下载包),之后瞬间完成。快捷方式:uv tool install tooluniverse → 然后直接使用 tu。
| 命令 | 功能 | 示例 |
|---|---|---|
tu status | 显示工具数量和顶级类别 | tu status |
tu list | 列出工具(模式:names, categories, basic, by_category, summary, custom) | tu list --mode basic --limit 20 |
tu find | 通过自然语言搜索(关键词评分,无需 API 密钥) | tu find 'protein structure analysis' |
tu grep | 文本/正则表达式模式搜索 | tu grep '^UniProt' --mode regex |
tu info | 显示工具参数和模式 | tu info PubMed_search_articles |
tu run | 执行工具 | tu run PubMed_search_articles '{"query": "CRISPR"}' |
tu test | 使用示例输入测试工具 | tu test UniProt_get_entry_by_accession |
tu build | 为 Coding API 生成类型化的 Python 包装器 | tu build --output ./my_tools |
tu serve | 启动 MCP stdio 服务器(与 uvx tooluniverse 相同) | tu serve |
输出标志(除 build/serve 外的大多数命令):--json(美观)或 --raw(紧凑,便于管道操作)。
继续到步骤 3(API 密钥)。
确保步骤 2 已完成。有关详细模式,请调用 tooluniverse-sdk 技能。
uv pip install tooluniverse
模式 1:直接导入(类型化,带自动补全):
from tooluniverse.tools import UniProt_get_entry_by_accession
result = UniProt_get_entry_by_accession(accession="P12345")
模式 2:属性访问(无需为每个工具导入):
from tooluniverse import ToolUniverse
tu = ToolUniverse()
tu.load_tools()
result = tu.tools.UniProt_get_entry_by_accession(accession="P12345")
模式 3:基于 JSON(动态,适用于流程):
result = tu.run({"name": "UniProt_get_entry_by_accession", "arguments": {"accession": "P12345"}})
生成类型化包装器:tu build(创建可导入的、带自动补全的 Python 模块)。
ToolUniverse 还包含 23 个 AI 驱动的智能体工具(ScientificTextSummarizer、HypothesisGenerator、ExperimentalDesignScorer、同行评审工具等)和 2 个代码执行器工具(python_code_executor、python_script_runner)。这些工具像其他任何工具一样调用——通过 tu.run() 或 execute_tool()。智能体工具需要 LLM API 密钥(例如 OPENAI_API_KEY)。
继续到步骤 3(API 密钥)。
确保步骤 2 已完成(uv --version 可用)。
配置文件帮助(如果用户似乎不熟悉):配置文件是存储设置的纯文本文件——就像应用的偏好设置列表。您无需理解格式;只需准确粘贴下面显示的内容。大多数应用都有一个设置按钮,可以为您打开该文件(见下表)。如果文件为空,请粘贴整个块。如果已有内容,智能体应帮助合并。
默认配置(大多数客户端相同):
{
"mcpServers": {
"tooluniverse": {
"command": "uvx",
"args": ["tooluniverse"],
"env": { "PYTHONIOENCODING": "utf-8" }
}
}
}
配置文件位置:
| 客户端 | 文件 | 访问方式 |
|---|---|---|
| Cursor | ~/.cursor/mcp.json | 设置 → MCP → 添加新的全局 MCP 服务器 |
| Claude Desktop | ~/Library/Application Support/Claude/claude_desktop_config.json | 设置 → 开发者 → 编辑配置 |
| Claude Code | ~/.claude.json 或 .mcp.json | claude mcp add 或直接编辑 |
| Windsurf | ~/.codeium/windsurf/mcp_config.json | MCP 锤子图标 → 配置 |
| Cline | cline_mcp_settings.json | Cline 面板 → MCP 服务器 → 配置 |
| Gemini CLI | ~/.gemini/settings.json | gemini mcp add 或直接编辑 |
| Trae | .trae/mcp.json | Ctrl+U → AI 管理 → MCP → 配置 |
不同格式:VS Code 使用 "servers" 键和 "type": "stdio"。Codex 使用 TOML。OpenCode 使用 "mcp" 键。有关这些,请参阅 references/mcp-configs.md。
继续到步骤 3(API 密钥)。
许多工具无需密钥即可工作,但有些工具能解锁强大功能。首先询问研究兴趣(AskQuestion):
映射到推荐的密钥(开始时 2-4 个)。逐一讲解:解释它能解锁什么,提供注册链接,等待密钥,添加到配置。
第 1 层(核心 — 推荐给大多数用户):
| 密钥 | 解锁功能 | 免费? | 注册 |
|---|---|---|---|
NCBI_API_KEY | PubMed(速率限制从 3 次/秒提升到 10 次/秒) | 是 | https://account.ncbi.nlm.nih.gov/settings/ |
NVIDIA_API_KEY | 16 个工具:AlphaFold2、对接、基因组学 | 是 | https://build.nvidia.com |
BIOGRID_API_KEY | 蛋白质相互作用查询 | 是 | https://webservice.thebiogrid.org/ |
FDA_API_KEY | FDA 不良事件、药品标签(速率从 240 次/分钟提升到 1000 次/分钟) | 是 | https://open.fda.gov/apis/authentication/ |
第 2 层(专业 — 基于兴趣):
| 密钥 | 解锁功能 | 注册 |
|---|---|---|
DISGENET_API_KEY | 基因-疾病关联 | https://disgenet.com/academic-apply |
OMIM_API_KEY | 孟德尔/罕见病 | https://omim.org/api |
ONCOKB_API_TOKEN | 精准肿瘤学 | https://www.oncokb.org/apiAccess |
UMLS_API_KEY | 医学术语 | https://uts.nlm.nih.gov/uts/ |
完整列表(包含所有层级)请参阅 API_KEYS_REFERENCE.md。
添加密钥:
聊天模式 — 添加到 MCP 配置中的 env 块:
"env": {
"PYTHONIOENCODING": "utf-8",
"NCBI_API_KEY": "your_key_here"
}
CLI — 设置环境变量:
export NCBI_API_KEY="your_key_here" # 当前会话
echo 'export NCBI_API_KEY="key"' >> ~/.zshrc # 跨会话持久化
SDK — 与 CLI 相同(导出或使用 .env 文件)。
不要只是告知 — 与用户一起操作。
聊天模式:要求用户重启应用。然后自己运行一个测试调用:
list_tools 或 grep_tools — 确认工具可见execute_tool("PubMed_search_articles", {"query": "CRISPR", "max_results": 1}) — 确认其工作CLI:一起运行:
tu status && tu find 'protein' && tu run PubMed_search_articles '{"query": "CRISPR", "max_results": 1}'
SDK:一起运行 SDK 设置中的 Python 代码片段。
如果出现问题:最常见原因:应用未重启、uv 不在 PATH 中(重新打开终端)、配置中的 JSON 语法错误。
技能是预构建的研究工作流,可将基本工具调用转变为专家级调查。
聊天模式用户:智能体应为用户运行此命令:
git clone --depth 1 https://github.com/mims-harvard/ToolUniverse.git /tmp/tu-skills
然后复制到客户端的技能目录:
| 客户端 | 命令 |
|---|---|
| Cursor | mkdir -p .cursor/skills && cp -r /tmp/tu-skills/skills/* .cursor/skills/ |
| Claude Code | mkdir -p .claude/skills && cp -r /tmp/tu-skills/skills/* .claude/skills/ |
| Windsurf | mkdir -p .windsurf/skills && cp -r /tmp/tu-skills/skills/* .windsurf/skills/ |
| Codex | mkdir -p .agents/skills && cp -r /tmp/tu-skills/skills/* .agents/skills/ |
| Gemini CLI | mkdir -p .gemini/skills && cp -r /tmp/tu-skills/skills/* .gemini/skills/ |
清理:rm -rf /tmp/tu-skills
技能会根据用户的问题自动激活。尝试:"Research the drug metformin" 或 "What does the literature say about CRISPR in cancer?"
CLI 用户:技能专为 AI 聊天智能体设计。请改用 tu find、tu info、tu run。对于完整的多步骤工作流,请使用聊天模式或构建 SDK 流程。
不要列出建议 — 与用户一起运行实时演示。
| 兴趣 | 首次查询 | 技能 |
|---|---|---|
| 文献 | "What does the literature say about CRISPR in cancer?" | literature-deep-research |
| 药物发现 | "Research the drug metformin" | drug-research |
| 蛋白质结构 | "Find protein structures for human EGFR" | protein-structure-retrieval |
| 基因组学 | "What genes are associated with type 2 diabetes?" | disease-research |
| 罕见病 | "Patient with progressive ataxia and oculomotor apraxia — differential diagnosis?" | rare-disease-diagnosis |
| 药物安全 | "What are the adverse events for pembrolizumab?" | pharmacovigilance |
| 通用 | "Research the drug aspirin" | drug-research |
运行演示 — 调用技能并显示真实结果。
聊天模式用户:
- "Research the drug [name]" — 完整的药物档案
- "Research [disease]" — 全面的疾病分析
- "What are the known targets of [drug]?" — 靶点情报
- "What does the literature say about [topic]?" — 深度文献综述
- "Find protein structures for [protein]" — 3D 结构
- "Is [variant] pathogenic?" — 变异解读
- "What drugs could be repurposed for [disease]?" — 药物再利用
- "What are the adverse events for [drug]?" — 安全性分析
- "Find clinical trials for [disease/drug]" — 试验匹配
- "What are the protein interactions for [gene]?" — 相互作用网络
- "What are the clinical guidelines for [condition]?" — 实践指南
- "Check drug interactions between [drug1] and [drug2]" — DDI 分析
只需自然提问 — 我会确定要搜索哪些数据库。
CLI 用户:
tu find 'drug safety' # 发现工具 tu info PubMed_search_articles # 检查参数 tu run PubMed_search_articles '{"query": "CRISPR cancer", "max_results": 3}' tu run UniProt_get_entry_by_accession '{"accession": "P12345"}' tu run FAERS_count_death_related_by_drug '{"drug_name": "metformin"}'
设置完成后,向智能体的记忆文件写入提醒,以便未来的会话知道 ToolUniverse 可用。
| 客户端 | 文件 |
|---|---|
| Cursor | .cursor/rules/tooluniverse.md |
| Claude Code | CLAUDE.md 或 .claude/CLAUDE.md(追加) |
| Windsurf | .windsurf/rules/tooluniverse.md |
| VS Code/Copilot | .github/copilot-instructions.md(追加) |
| Codex / OpenCode | AGENTS.md(追加) |
| Gemini CLI | GEMINI.md(追加) |
| Cline | .clinerules/tooluniverse.md |
| Trae | .trae/rules/tooluniverse.md |
内容:
# ToolUniverse
For any scientific research question (drugs, genes, proteins, diseases, literature, clinical trials, etc.), invoke the `tooluniverse` skill — it routes to specialized research skills and 1,200+ database tools.
追加(不要覆盖)。首先检查是否存在现有部分。征求用户许可。
如果为团队或共享项目设置 ToolUniverse:
共享 API 密钥:在项目根目录创建一个包含所有密钥的 .env 文件。大多数客户端以及 CLI/SDK 会自动从 .env 中获取密钥:
NCBI_API_KEY=your_shared_key
NVIDIA_API_KEY=your_shared_key
项目级 MCP 配置(以便所有团队成员自动获得 ToolUniverse):
.cursor/mcp.json.mcp.json.vscode/mcp.json项目级技能:将技能安装到项目中(例如,.cursor/skills/),以便所有团队成员共享。
团队范围升级:每个团队成员运行 uv cache clean tooluniverse 并重启其应用。要固定特定版本,请在 MCP 配置中使用 "args": ["tooluniverse==X.Y.Z"]。
| 问题 | 解决方法 |
|---|---|
requires-python >= 3.10 | uv python install 3.12 |
uvx: command not found | 运行步骤 2 中的安装脚本,重启终端 |
| 上下文窗口溢出 | 验证是否使用 uvx tooluniverse(精简模式是默认的) |
ModuleNotFoundError | uv pip install tooluniverse[all] |
| MCP 服务器无法启动 | 测试:在终端中运行 uvx tooluniverse。检查 JSON 语法。 |
| API 密钥 401/403 | 检查 env 块中的密钥,重启应用,验证密钥名称 |
| 需要升级 | uv cache clean tooluniverse 然后重启应用 |
仍然卡住?GitHub issues 或发送邮件至 Shanghua Gao。
uvx tooluniverse — 自动安装,精简模式uv cache clean tooluniverse + 重启每周安装量
170
仓库
GitHub 星标数
1.2K
首次出现
2026年2月8日
安全审计
安装于
codex163
opencode162
gemini-cli161
github-copilot159
kimi-cli154
amp154
Guide the user step-by-step through setting up ToolUniverse.
ToolUniverse has 1200+ tools. The tooluniverse command enables compact mode automatically, exposing only 5 core MCP tools (list_tools, grep_tools, get_tool_info, execute_tool, find_tools) while keeping all tools accessible via execute_tool.
Always explain first, in plain language:
ToolUniverse is free, open-source software connecting to 2,000+ scientific databases (PubMed, UniProt, ChEMBL, FAERS, ClinicalTrials.gov, etc.). Instead of visiting each website, you search from one place. Think of it like a universal remote for scientific databases.
Why AI assistants? The AI reads your question, figures out which databases to search, runs queries, and summarizes results. You just ask your question.
Present using AskQuestion:
| Mode | What it means | Who it's for |
|---|---|---|
| Chat mode | Ask questions to an AI assistant. No coding. | Most researchers. |
| Command line | Type short commands in Terminal. | Quick tests. Terminal-comfortable users. |
| Python code | Write scripts for automated pipelines. | Programmers. |
Options: "I want to ask questions" → Chat mode | "Quick try" → CLI | "I write Python" → SDK | "I don't know" → Recommend Chat mode
If Chat mode , ask which app (AskQuestion): Cursor, Claude Desktop, VS Code/Copilot, Windsurf, Claude Code, Gemini CLI, Codex, Cline/Trae/Antigravity/OpenCode. "I don't have any" → Recommend Claude Desktop.
Only prerequisite: uv (manages everything else automatically).
Terminal help (if needed): Mac: Cmd+Space → "Terminal" → Enter. Windows: Win key → "PowerShell" → Enter.
curl -LsSf https://astral.sh/uv/install.sh | sh
(This is a safe, standard command that downloads and installs uv, a small package manager. It's widely used by Python developers. Close and reopen your terminal after it finishes.)
Verify: uv --version
Make sure Step 2 is done, then try:
uvx --from tooluniverse tu status # How many tools?
uvx --from tooluniverse tu find 'drug safety' # Search by topic
uvx --from tooluniverse tu info FAERS_count_death_related_by_drug # See params
uvx --from tooluniverse tu run FAERS_count_death_related_by_drug '{"drug_name": "metformin"}'
First run takes ~30s (downloads package), then instant. Shortcut : uv tool install tooluniverse → then just use tu directly.
| Command | What it does | Example |
|---|---|---|
tu status | Show tool count and top categories | tu status |
tu list | List tools (modes: names, categories, basic, by_category, summary, custom) | tu list --mode basic --limit 20 |
tu find | Search by natural language (keyword scoring, no API key needed) | tu find 'protein structure analysis' |
Output flags (most commands except build/serve): --json (pretty) or --raw (compact, pipe-friendly).
Continue to Step 3 (API Keys).
Make sure Step 2 is done. For detailed patterns, invoke the tooluniverse-sdk skill.
uv pip install tooluniverse
Pattern 1: Direct import (typed, with autocomplete):
from tooluniverse.tools import UniProt_get_entry_by_accession
result = UniProt_get_entry_by_accession(accession="P12345")
Pattern 2: Attribute access (no import needed per tool):
from tooluniverse import ToolUniverse
tu = ToolUniverse()
tu.load_tools()
result = tu.tools.UniProt_get_entry_by_accession(accession="P12345")
Pattern 3: JSON-based (dynamic, for pipelines):
result = tu.run({"name": "UniProt_get_entry_by_accession", "arguments": {"accession": "P12345"}})
Generate typed wrappers: tu build (creates importable Python modules with autocomplete).
ToolUniverse also includes 23 AI-powered agentic tools (ScientificTextSummarizer, HypothesisGenerator, ExperimentalDesignScorer, peer-review tools, etc.) and 2 code executor tools (python_code_executor, python_script_runner). These are called like any other tool — via tu.run() or execute_tool(). Agentic tools require an LLM API key (e.g., OPENAI_API_KEY).
Continue to Step 3 (API Keys).
Make sure Step 2 is done (uv --version works).
Config file help (if user seems unfamiliar): Config files are plain text that store settings — like a preference list for the app. You don't need to understand the format; just paste exactly what's shown below. Most apps have a Settings button that opens the file for you (see table). If the file is empty, paste the entire block. If it already has content, the agent should help merge it.
Default config (same for most clients):
{
"mcpServers": {
"tooluniverse": {
"command": "uvx",
"args": ["tooluniverse"],
"env": { "PYTHONIOENCODING": "utf-8" }
}
}
}
Config file locations:
| Client | File | How to Access |
|---|---|---|
| Cursor | ~/.cursor/mcp.json | Settings → MCP → Add new global MCP server |
| Claude Desktop | ~/Library/Application Support/Claude/claude_desktop_config.json | Settings → Developer → Edit Config |
| Claude Code | ~/.claude.json or .mcp.json | claude mcp add or edit directly |
| Windsurf | ~/.codeium/windsurf/mcp_config.json |
Different formats : VS Code uses "servers" key with "type": "stdio". Codex uses TOML. OpenCode uses "mcp" key. See references/mcp-configs.md for these.
Continue to Step 3 (API Keys).
Many tools work without keys, but some unlock powerful features. Ask research interests first (AskQuestion):
Map to recommended keys (2-4 to start). Walk through one at a time : explain what it unlocks, give registration link, wait for key, add to config.
Tier 1 (Core — recommend for most users):
| Key | Unlocks | Free? | Registration |
|---|---|---|---|
NCBI_API_KEY | PubMed (rate limit 3→10/s) | Yes | https://account.ncbi.nlm.nih.gov/settings/ |
NVIDIA_API_KEY | 16 tools: AlphaFold2, docking, genomics | Yes | https://build.nvidia.com |
BIOGRID_API_KEY | Protein interaction queries | Yes | https://webservice.thebiogrid.org/ |
FDA_API_KEY |
Tier 2 (Specialized — based on interests):
| Key | Unlocks | Registration |
|---|---|---|
DISGENET_API_KEY | Gene-disease associations | https://disgenet.com/academic-apply |
OMIM_API_KEY | Mendelian/rare disease | https://omim.org/api |
ONCOKB_API_TOKEN | Precision oncology | https://www.oncokb.org/apiAccess |
UMLS_API_KEY | Medical terminology | https://uts.nlm.nih.gov/uts/ |
See API_KEYS_REFERENCE.md for the complete list with all tiers.
Adding keys:
Chat mode — add to env block in MCP config:
"env": {
"PYTHONIOENCODING": "utf-8",
"NCBI_API_KEY": "your_key_here"
}
CLI — set environment variables:
export NCBI_API_KEY="your_key_here" # Current session
echo 'export NCBI_API_KEY="key"' >> ~/.zshrc # Persist across sessions
SDK — same as CLI (export or .env file).
Don't just tell — do it WITH the user.
Chat mode : Ask user to restart app. Then run a test call yourself:
list_tools or grep_tools with "PubMed" — confirm tools visibleexecute_tool("PubMed_search_articles", {"query": "CRISPR", "max_results": 1}) — confirm it worksCLI : Run together:
tu status && tu find 'protein' && tu run PubMed_search_articles '{"query": "CRISPR", "max_results": 1}'
SDK : Run the Python snippet from SDK Setup together.
If issues : Most common: app not restarted, uv not in PATH (reopen terminal), JSON syntax error in config.
Skills are pre-built research workflows that turn basic tool calls into expert investigations.
Chat mode users : The agent should run this for the user:
git clone --depth 1 https://github.com/mims-harvard/ToolUniverse.git /tmp/tu-skills
Then copy to client's skill directory:
| Client | Command |
|---|---|
| Cursor | mkdir -p .cursor/skills && cp -r /tmp/tu-skills/skills/* .cursor/skills/ |
| Claude Code | mkdir -p .claude/skills && cp -r /tmp/tu-skills/skills/* .claude/skills/ |
| Windsurf | mkdir -p .windsurf/skills && cp -r /tmp/tu-skills/skills/* .windsurf/skills/ |
| Codex | mkdir -p .agents/skills && cp -r /tmp/tu-skills/skills/* .agents/skills/ |
| Gemini CLI | mkdir -p .gemini/skills && cp -r /tmp/tu-skills/skills/* .gemini/skills/ |
Clean up: rm -rf /tmp/tu-skills
Skills activate automatically based on user's question. Try: "Research the drug metformin" or "What does the literature say about CRISPR in cancer?"
CLI users : Skills are designed for AI chat agents. Use tu find, tu info, tu run instead. For full multi-step workflows, use Chat mode or build SDK pipelines.
Don't list suggestions — run a live demo WITH the user.
| Interest | First query | Skill |
|---|---|---|
| Literature | "What does the literature say about CRISPR in cancer?" | literature-deep-research |
| Drug discovery | "Research the drug metformin" | drug-research |
| Protein structure | "Find protein structures for human EGFR" | protein-structure-retrieval |
| Genomics | "What genes are associated with type 2 diabetes?" | disease-research |
| Rare diseases | "Patient with progressive ataxia and oculomotor apraxia — differential diagnosis?" | rare-disease-diagnosis |
| Drug safety | "What are the adverse events for pembrolizumab?" | pharmacovigilance |
| General | "Research the drug aspirin" | drug-research |
Run the demo — invoke the skill and show real results.
Chat mode users:
- "Research the drug [name]" — full drug profile
- "Research [disease]" — comprehensive disease analysis
- "What are the known targets of [drug]?" — target intelligence
- "What does the literature say about [topic]?" — deep literature review
- "Find protein structures for [protein]" — 3D structures
- "Is [variant] pathogenic?" — variant interpretation
- "What drugs could be repurposed for [disease]?" — repurposing
- "What are the adverse events for [drug]?" — safety analysis
- "Find clinical trials for [disease/drug]" — trial matching
- "What are the protein interactions for [gene]?" — interaction network
- "What are the clinical guidelines for [condition]?" — practice guidelines
- "Check drug interactions between [drug1] and [drug2]" — DDI analysis
Just ask naturally — I'll figure out which databases to search.
CLI users:
tu find 'drug safety' # Discover tools tu info PubMed_search_articles # Check parameters tu run PubMed_search_articles '{"query": "CRISPR cancer", "max_results": 3}' tu run UniProt_get_entry_by_accession '{"accession": "P12345"}' tu run FAERS_count_death_related_by_drug '{"drug_name": "metformin"}'
After setup, write a reminder to the agent's memory file so future sessions know ToolUniverse is available.
| Client | File |
|---|---|
| Cursor | .cursor/rules/tooluniverse.md |
| Claude Code | CLAUDE.md or .claude/CLAUDE.md (append) |
| Windsurf | .windsurf/rules/tooluniverse.md |
| VS Code/Copilot | .github/copilot-instructions.md (append) |
| Codex / OpenCode | AGENTS.md (append) |
| Gemini CLI |
Content:
# ToolUniverse
For any scientific research question (drugs, genes, proteins, diseases, literature, clinical trials, etc.), invoke the `tooluniverse` skill — it routes to specialized research skills and 1,200+ database tools.
Append (don't overwrite). Check for existing section first. Ask user permission.
If setting up ToolUniverse for a team or shared project:
Shared API keys : Create a .env file at the project root with all keys. Most clients and the CLI/SDK will pick up keys from .env automatically:
NCBI_API_KEY=your_shared_key
NVIDIA_API_KEY=your_shared_key
Project-level MCP config (so all team members get ToolUniverse automatically):
.cursor/mcp.json in project root.mcp.json in project root.vscode/mcp.json in project rootProject-level skills : Install skills into the project (e.g., .cursor/skills/) so all team members share them.
Team-wide upgrade : Each team member runs uv cache clean tooluniverse and restarts their app. To pin a specific version, use "args": ["tooluniverse==X.Y.Z"] in the MCP config.
| Issue | Fix |
|---|---|
requires-python >= 3.10 | uv python install 3.12 |
uvx: command not found | Run install script from Step 2, restart terminal |
| Context window overflow | Verify using uvx tooluniverse (compact mode is default) |
ModuleNotFoundError | uv pip install tooluniverse[all] |
| MCP server won't start | Test: uvx tooluniverse in terminal. Check JSON syntax. |
Still stuck? GitHub issues or email Shanghua Gao.
uvx tooluniverse — auto-installs, compact modeuv cache clean tooluniverse + restartWeekly Installs
170
Repository
GitHub Stars
1.2K
First Seen
Feb 8, 2026
Security Audits
Gen Agent Trust HubPassSocketWarnSnykFail
Installed on
codex163
opencode162
gemini-cli161
github-copilot159
kimi-cli154
amp154
Lark CLI妙记查询工具:快速获取飞书妙记元信息(标题、封面、时长)
32,800 周安装
ByteRover CLI - 上下文工程平台,为AI编码智能体自动管理项目知识库
154 周安装
嵌入式系统开发指南:RTOS、裸机固件、嵌入式Linux与微控制器编程实战
155 周安装
Markdown转EPUB转换器 - 专业电子书制作工具,支持代码高亮与表格样式
158 周安装
化学品安全与毒理学评估工具 | AI预测毒理学、ADMET分析、基因互作与监管安全数据整合
156 周安装
gentle-teaching 温和教学框架:AI辅助学习指南,培养独立解决问题能力
156 周安装
iOS应用可发现性优化指南:通过App Intents、App Shortcuts和Core Spotlight提升应用曝光
157 周安装
tu grep| Text/regex pattern search |
tu grep '^UniProt' --mode regex |
tu info | Show tool parameters and schema | tu info PubMed_search_articles |
tu run | Execute a tool | tu run PubMed_search_articles '{"query": "CRISPR"}' |
tu test | Test a tool with its example inputs | tu test UniProt_get_entry_by_accession |
tu build | Generate typed Python wrappers for Coding API | tu build --output ./my_tools |
tu serve | Start MCP stdio server (same as uvx tooluniverse) | tu serve |
| MCP hammer icon → Configure |
| Cline | cline_mcp_settings.json | Cline panel → MCP Servers → Configure |
| Gemini CLI | ~/.gemini/settings.json | gemini mcp add or edit directly |
| Trae | .trae/mcp.json | Ctrl+U → AI Management → MCP → Configure |
| FDA adverse events, drug labels (rate 240→1000/min) |
| Yes |
| https://open.fda.gov/apis/authentication/ |
GEMINI.md (append) |
| Cline | .clinerules/tooluniverse.md |
| Trae | .trae/rules/tooluniverse.md |
| API key 401/403 | Check key in env block, restart app, verify key name |
| Upgrade needed | uv cache clean tooluniverse then restart app |