Agent Skills by centaurioun
Showing 25 vetted skills indexed across 1 GitHub repositories.
tooluniverse
Route scientific queries to specialized ToolUniverse skills via keyword extraction and routing tables.
tooluniverse-custom-tool
Register and call private REST APIs or local Python tools in ToolUniverse.
tooluniverse-statistical-modeling
Fit biomedical regression models and compute effect estimates with statsmodels and lifelines.
setup-tooluniverse
Install and configure ToolUniverse for MCP chat, CLI, and Python SDK usage.
devtu-docs-quality
Detect and correct incorrect, outdated, and redundant documentation in ToolUniverse repositories.
devtu-self-evolve
Orchestrate the ToolUniverse self-improvement cycle from discovery to pull requests.
create-tooluniverse-skill
Generate tested ToolUniverse skills with templates and fallback strategies.
devtu-code-optimization
Applies code-quality patterns to Python ToolUniverse projects.
devtu-create-tool
Create ToolUniverse-compatible Python tool classes with JSON configurations and tests.
devtu-optimize-descriptions
Optimize ToolUniverse JSON tool descriptions with prerequisites and filter syntax.
devtu-optimize-skills
Optimize ToolUniverse research skills to produce evidence-graded narrative reports with source attribution.
tooluniverse-install-skills
Detect missing ToolUniverse skills and install them via sparse git checkout.
devtu-fix-tool
Diagnose and fix failing ToolUniverse tools and their test failures.
tooluniverse-rare-disease-diagnosis
Map phenotypes to HPO terms and rank rare disease diagnoses.
tooluniverse-literature-deep-research
Conduct collision-aware literature searches and evidence-graded syntheses for academic topics.
devtu-github
Automate GitHub pushes with pre-commit hooks, pytest, and rebase workflows.
devtu-auto-discover-apis
Automate discovery, validation, and integration of life-science APIs into ToolUniverse.
tooluniverse-[domain-name]
Aggregate domain queries across biomedical databases into phased Markdown reports.
run-report-writer
Populate run report templates with execution evidence from logs and manifests.
authority-resolution
Classify file paths into canonical, manuscript, archival, or reference authority domains.
claim-confidence-registry
Map manuscript claims to evidence cells and assign confidence classes.
comparison-report-writer
Classify discrepancies between baseline and regenerated outputs into a comparison report.
runtime-transform-checks
Validate runtime variable transformations in analysis scripts against workspace semantic rules.
small-n-stats-guardrails
Validate small-sample study results with statistical guardrails and correction checks.