ToolUniverse is a collection of tools, interfaces, and supporting components for building AI systems that perform scientific work. It is for developers creating AI scientist agents that use APIs, databases, machine-learning tools, and domain-specific utilities. The catalogue includes skills, commands, an MCP server, an agent, and a hook for working with the ecosystem.
Getting it into your agent
One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.
npx skills add mims-harvard/ToolUniverse --skill tooluniverse-publish-toolgit clone --depth 1 https://github.com/mims-harvard/ToolUniverseWrote 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/skills/mims-harvard/tooluniverse/tooluniverse-publish-tool)<a href="https://agentmods.dev/skills/mims-harvard/tooluniverse/tooluniverse-publish-tool"><img src="https://agentmods.dev/badge/skills/mims-harvard/tooluniverse/tooluniverse-publish-tool/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/skills/mims-harvard/tooluniverse/tooluniverse-publish-tool"><img src="https://agentmods.dev/badge/skills/mims-harvard/tooluniverse/tooluniverse-publish-tool.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 | $0.00103 | $0.01200 |
| Opus 5.5 | $0.00041 | $0.00480 |
| Sonnet 5.5 | $0.00021 | $0.00240 |
| Haiku 4.5 | $0.00010 | $0.00120 |
Grade A, and why
tooluniverse-publish-tool 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 yesterday.
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 — 65 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Publish one tool to ToolUniverse
Publishing puts one tool from a computer the user already shares into ToolUniverse, with a page in Discover if they make it public. People can run that one tool; they cannot see the computer's other tools or anything else on it.
The person you are helping is usually a scientist, not a programmer. You do the preparation; the final choice of who can use it, and the Publish click, are the user's on the website. Never ask for passwords, keys or tokens in the chat.
1. Make sure the tool is shared first
Publishing needs the tool to be running on a computer connected to ToolUniverse.
- If you are connected to the hosted ToolUniverse (you have
list_remote_tools), call it and look for the user's computer ("your computer ...") and the tool. - Otherwise ask the user whether https://connect.aiscientist.tools/remote-servers shows the computer online with the tool listed.
If it is not shared yet, do that first with the skill tooluniverse-share-computer, then come back.
2. Decide who may use it
Ask in plain words and explain the choice:
- Only me: a private tool for their own assistant. Not listed anywhere.
- Specific people: only colleagues they add by their ToolUniverse account email.
- Public: listed in Discover; anyone signed in to ToolUniverse can run it. Confirm the user has the right to share the model, data and results with everyone, and that the computer can handle the load, before choosing this.
To give a few colleagues every tool on that computer instead of one tool, publishing is not needed: use a share code (see tooluniverse-share-computer, step 7).
3. Prepare it, then let the user publish
If you have the prepare_tool_publication tool (the hosted ToolUniverse connection), do the preparation yourself:
- Write a clear title, one or two sentences on what it does (what goes in, what comes out), and a few search keywords, from what you know about the tool. Show them to the user and adjust.
- Pick a small, realistic test input that matches the tool's inputs (see
list_remote_toolsfor its input schema). - Call
prepare_tool_publicationwithcomputer(its name aslist_remote_toolsshows it),tool,name,description,keywordsandtest_arguments. It creates (or reuses) the private draft and runs a real test call. - If
test_passedis false, readtest_error, fix the cause on the computer (the tool itself, or the computer went offline) and call it again. - When it passes, give the user the
review_linkand say exactly this: open it, choose who can use it (Choose who can use it: Only me, Specific people with their account emails, or Public), then click Publish. You cannot publish for them: that choice is theirs, on purpose.
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.
- yesterday Changed 5a1a3adf7734
- 7d ago Changed · +8 lines · +5 tokens per session 16db7c45da29
- 8d ago First seen · 57 lines · 98 tokens per session scan A 7f1b5f22274c
tooluniverse-publish-tool is a skill published in the GitHub repository mims-harvard/ToolUniverse (1,744 stars, last pushed yesterday), licensed Apache-2.0. It adds 103 tokens to every session and 1,200 once invoked, about $0.0004 per session on Opus 5.5. 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 2026-10-09.
Other skills, from other repositories
ai-agent-tool-routing
Use when designing or reviewing an AI agent tool registry or MCP router: document parameter schemas, caller authorization, execution limits, replay-safe retries and verification evidence.
nanoresearch-writing
Draft a LaTeX research paper from all previous stage outputs.
construct-toy-examples
Generate and analyze simpler examples that satisfy both the assumptions and the conclusion of a theorem statement or subgoal. Use when you are stuck in reasoning and need simpler examples to regain traction, or when you want to see where the assumptions take effect and gain intuition.
obtain-immediate-conclusions
Derive immediate mathematical consequences from a theorem statement or subgoal. Use when starting a new problem, branch, or subgoal, or when cheap progress or a cleaner reformulation is needed before deeper proof search.
clinical-research-analysis-framework
Guided workflow for statistical and ML analysis of clinical data. Use when planning or executing research analyses on MIMIC, eICU, or similar EHR data. Ensures methodological rigor through structured consultation, assumption checking, and stepwise execution with audit trails.
apache-iv-score
Calculate APACHE IV (Acute Physiology and Chronic Health Evaluation IV) score for ICU mortality prediction. Use for severity assessment, hospital mortality prediction, ICU benchmarking, or case-mix adjustment. eICU has pre-computed scores; MIMIC-IV requires custom implementation with diagnosis mapping challenges.