What it does\nToolUniverse Gateway exposes a standardized tu.run() interface so agents can call hundreds of research tools (PubChem, UniProt, GWAS workflows, proteomics pipelines, and more) and receive JSON-formatted results. This skill wraps the ToolUniverse client with helper scripts for executing specific tools, listing available tools, and formatting outputs for the scienceclaw artifact system. Concrete use cases include retrieving protein entries, running compound/property lookups, and launching domain-specific research workflows like GWAS or drug-repurposing.\n\n## When to use it\nUse this skill when the agent must query scientific databases, run bioinformatics or chemical analysis tools, or orchestrate multi-step research pipelines (e.g., retrieve UniProt entries, run PubChem property lookups, or execute standardized GWAS workflows). Trigger phrases include: "run ToolUniverse", "search UniProt", "run PubChem_query", or any request for programmatic scientific tool access.\n\n## What's included\n- Scripts: tooluniverse_run.py (run tools), tooluniverse_list.py (discover/search tools), setup helper scripts\n- References: tool list and example workflows embedded in SKILL.md\n- Instructions: examples show how to run tools, format args as JSON, and handle caching and output formats. Notes cover API-key usage for downstream tools.\n\n## Compatible agents\nCompatible with agents that can run Python helper scripts and shell commands (e.g., Copilot/Code agents, Cursor, Claude Code). The skill assumes access to environment variables for upstream API keys and a Python runtime.
Not yet audited
This skill has not been reviewed by our automated audit pipeline yet.
