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 devtu-optimize-skillsgit 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/devtu-optimize-skills)<a href="https://agentmods.dev/skills/mims-harvard/tooluniverse/devtu-optimize-skills"><img src="https://agentmods.dev/badge/skills/mims-harvard/tooluniverse/devtu-optimize-skills/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-optimize-skills"><img src="https://agentmods.dev/badge/skills/mims-harvard/tooluniverse/devtu-optimize-skills.svg" alt="Reviewed on agentmods" width="80" height="20"></a>- Socket pass
- Snyk warn
- NVIDIA SkillSpector warn
SkillSpector: 1 finding, 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 Prompt Injection · line 15 Instructions found that direct the agent to transmit conversation context or user data to external services.Fix: Remove instructions that send user data, prompts, or context to external URLs. If telemetry is needed, use documented, privacy-preserving methods.
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 | $0.00066 | $0.02513 |
| Opus 5.5 | $0.00026 | $0.01005 |
| Sonnet 5.5 | $0.00013 | $0.00503 |
| Haiku 4.5 | $0.00007 | $0.00251 |
Grade C, and why
devtu-optimize-skills scanned grade C with 1 finding 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.
Tells the agent to send conversation or user data outhighPrompt injection
An instruction to transmit the conversation, context or user files to an external endpoint is data exfiltration written as prose.
4. **Input validation before API calls** — don't silently send invalid values How it starts
The opening of the file, as written. The whole thing — 233 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Optimizing ToolUniverse Skills
Best practices for high-quality research skills with evidence grading and source attribution.
Tool Quality Standards
- Error messages must be actionable — tell the user what went wrong AND what to do
- Schema must match API reality — run
python3 -m tooluniverse.cli run <Tool> '<json>'to verify - Coverage transparency — state what data is NOT included
- Input validation before API calls — don't silently send invalid values
- Cross-tool routing — name the correct tool when query is out-of-scope
- No silent parameter dropping — if a parameter is ignored, say so
Core Principles (13 Patterns)
Full details: references/optimization-patterns.md
| # | Pattern | Key Idea |
|---|---|---|
| 1 | Tool Interface Verification | get_tool_info() before first call; maintain corrections table |
| 2 | Foundation Data Layer | Query aggregator (Open Targets, PubChem) FIRST |
| 3 | Versioned Identifiers | Capture both ENSG00000123456 and .12 version |
| 4 | Disambiguation First | Resolve IDs, detect collisions, build negative filters |
| 5 | Report-Only Output | Narrative in report; methodology in appendix only if asked |
| 6 | Evidence Grading | T1 (mechanistic) → T2 (functional) → T3 (association) → T4 (mention) |
| 7 | Quantified Completeness | Numeric minimums per section (>=20 PPIs, top 10 tissues) |
| 8 | Mandatory Checklist | All sections exist, even if "Limited evidence" |
| 9 | Aggregated Data Gaps | Single section consolidating all missing data |
| 10 | Query Strategy | High-precision seeds → citation expansion → collision-filtered broad |
| 11 | Tool Failure Handling | Primary → Fallback 1 → Fallback 2 → document unavailable |
| 12 | Scalable Output | Narrative report + JSON/CSV bibliography |
| 13 | Synthesis Sections | Biological model + testable hypotheses, not just paper lists |
Optimized Skill Workflow
Phase -1: Tool Verification (check params)
Phase 0: Foundation Data (aggregator query)
Phase 1: Disambiguation (IDs, collisions, baseline)
Phase 2: Specialized Queries (fill gaps)
Phase 3: Report Synthesis (evidence-graded narrative)
What ships with it
4 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 · 233 lines · 66 tokens per session scan C 6e46e3d2bfd4
devtu-optimize-skills is a skill published in the GitHub repository mims-harvard/ToolUniverse (1,739 stars, last pushed today), licensed Apache-2.0. It adds 66 tokens to every session and 2,513 once invoked, about $0.0003 per session on Opus 5.5. A static security scan graded it C with 1 finding (tells the agent to send conversation or user data out). No closer match exists in the catalogue, so it is treated as the original; first seen 2026-09-03.
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.
agent-engineering-expert
Build LLM agents that use tools safely: tool design, the agent loop, memory, MCP servers, multi-agent orchestration, sandboxing and prompt-injection defence. Use when the user mentions AI agents, tool use or function calling, MCP or Model Context Protocol, autonomous workflows, multi-agent systems, LangChain or…
external-repo-setup
Protocol for GPU agents to clone external GitHub repos, install their dependencies, download pretrained weights, and integrate them as featurizers or models inside a focus-area train.py. Use this when a proposal references a GitHub repo or pretrained checkpoint that is not already present in the focus-area workspace.
multi-agent-focus
Generic skill for self-organizing multi-agent teams that collaborate on an optimization problem. Agents discuss dimensions, form teams, run experiments, and adapt when stagnating. Uses AnonAPI posts for discussion and workspaces for shared state.
agentic-patterns
Fundamental patterns for effective agentic behavior. Teaches decomposition, tool orchestration, error recovery, context management, quality self-assessment, and knowing when to stop. Model-agnostic principles that make any agent more effective regardless of domain. Activate on: "how should I structure this agent"…
ai-agent-tool-routing
Autonomous AI agent tool router register: schema registration, runtime parameter coercion, idempotency keys, and error recovery policies.