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
It runs from inside its repository, so the clone comes first — what it calls does not travel with the file alone.
git clone --depth 1 https://github.com/mims-harvard/ToolUniversenpx agentmods add skills/mims-harvard/tooluniverse/devtu-self-evolveWrote 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/devtu-self-evolve)<a href="https://agentmods.dev/skills/mims-harvard/tooluniverse/devtu-self-evolve"><img src="https://agentmods.dev/badge/skills/mims-harvard/tooluniverse/devtu-self-evolve/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/devtu-self-evolve"><img src="https://agentmods.dev/badge/skills/mims-harvard/tooluniverse/devtu-self-evolve.svg" alt="Reviewed on agentmods" width="80" height="20"></a>- NVIDIA SkillSpector warn
SkillSpector: 2 findings, up to high
These are SkillSpector’s own severities. On a checked sample its high-severity flags on skills were ~96% false positives — a documented command, a public API, a “never do X” rule — so we show them as a caution to read, not a verdict. Why →
- high Rogue Agent · line 2 Skill modifies its own code, configuration, or behavior at runtime. Self-modification enables an agent to escalate privileges, disable safety constraints, or install persistent backdoors.Fix: Prevent the skill from modifying its own code, SKILL.md, or configuration files. Treat skill files as read-only at runtime.
- high Tool Misuse · line 163 Tool parameters are crafted to achieve unintended or unsafe behavior. Parameter abuse can bypass intended safety checks (e.g. shell=True, --force, dangerous glob patterns).Fix: Validate all tool parameters against an allowlist. Reject dangerous parameter values (shell=True, --force, -rf /) and use safe defaults.
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 | $0.00082 | $0.02222 |
| Opus 5.5 | $0.00033 | $0.00889 |
| Sonnet 5.5 | $0.00016 | $0.00444 |
| Haiku 4.5 | $0.00008 | $0.00222 |
Grade A, and why
devtu-self-evolve 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 1mo 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 — 212 lines — stays where its author put it; the contents beside it link to each section on GitHub.
ToolUniverse Self-Evolution Orchestrator
Coordinates the full development lifecycle by dispatching to specialized devtu skills.
The Cycle
Discover → Create → Test → Fix → Optimize → Ship → Repeat
Each phase maps to a dedicated skill:
| Phase | Skill | What it does |
|---|---|---|
| Discover | devtu-auto-discover-apis |
Gap analysis, web search for APIs, batch discovery |
| Create | devtu-create-tool |
Build tool class + JSON config + test examples |
| Test | (this skill) | Launch researcher persona agents to find issues |
| Fix | devtu-fix-tool |
Diagnose failures, implement fixes, validate |
| Optimize | devtu-optimize-skills |
Improve skill reports, evidence handling, UX |
| Optimize | devtu-optimize-descriptions |
Improve tool JSON descriptions for clarity |
| Docs | devtu-docs-quality |
Validate documentation accuracy |
| Ship | devtu-github |
Branch, commit, push, create PR |
Quick Start
Pick an entry point based on what's needed:
- "Run a test round" → jump to Testing Phase
- "Expand coverage" → invoke
Skill(skill="devtu-auto-discover-apis") - "Create a new tool" → invoke
Skill(skill="devtu-create-tool") - "Fix a broken tool" → invoke
Skill(skill="devtu-fix-tool") - "Improve skills" → invoke
Skill(skill="devtu-optimize-skills") - "Full cycle" → follow all phases below in order
Phase 1: Discovery (optional)
Invoke Skill(skill="devtu-auto-discover-apis") to:
- Run gap analysis on current tool categories
- Search for life science APIs in underrepresented domains
- Score and prioritize APIs by coverage, reliability, documentation
Phase 2: Tool Creation (optional)
Invoke Skill(skill="devtu-create-tool") for each new API:
- Create Python tool class implementing the API
- Create JSON config with parameters, descriptions, test examples
- Register in
_lazy_registry_static.pyanddefault_config.py - Validate:
python -m tooluniverse.cli test <ToolName>
What ships with it
2 files beside SKILL.md in the same directory: the scripts, references and assets a skill reads on demand. Not counted in the per-session cost; read them before you install if any of them is executable.
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.
- 1mo ago First seen · 212 lines · 82 tokens per session scan A 258cb773d59e
devtu-self-evolve is a skill published in the GitHub repository mims-harvard/ToolUniverse (1,739 stars, last pushed today), licensed Apache-2.0. It adds 82 tokens to every session and 2,222 once invoked, about $0.0003 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 2026-09-03.
Other skills, from other repositories
style-refresh
Refresh CocoStyle conventions from project evidence.
style-show
Show CocoStyle conventions for the current project.
product-management
Product management expertise for product strategy, roadmap planning, feature prioritization (RICE, ICE, MoSCoW), customer research, A/B testing, product analytics, and product-market fit. Use when building product roadmaps, prioritizing features, or defining product strategy.
yak-worktree-workflow
Use when starting work on a yak - sets up an isolated git worktree, reads yak context, and guides the full cycle from claiming through merge and cleanup.
curate-collection
Build and maintain a library collection through acquisitions, weeding (deaccessioning), collection assessment, reader advisory, and interlibrary loan coordination. Covers selection criteria, collection development policies, the CREW/MUSTIE method for weeding, usage analysis, and responsive collection management. Use…
work-on-ticket
Fetches Jira ticket details, creates an appropriately named branch, and initiates the task planning workflow. Use when the user says "work on [TICKETID]" or similar phrases.