{ "markdown": "# SciAgent-Skills\n\n

\n \"199\n \"BixBench\n \"CC-BY-4.0\"\n \"GitHub\n

\n\n> **Turn your AI coding agent into a life sciences expert** — 199 bioinformatics skills for Claude Code covering RNA-seq, single-cell analysis, genomics, proteomics, drug discovery, and more. Boosted [BixBench](https://github.com/Future-House/BixBench) from 65% to 92%. Open source.\n\n**SciAgent-Skills** is the largest open-source skill library for scientific AI agents. It equips [Claude Code](https://docs.anthropic.com/en/docs/claude-code) (and any markdown-compatible agent) with domain-specific knowledge for computational biology, bioinformatics, cheminformatics, and biostatistics — no fine-tuning required, just plug in and analyze.\n\n**Keywords:** bioinformatics AI agent, Claude Code skills, scientific computing, RNA-seq analysis, single-cell RNA-seq, drug discovery pipeline, protein structure prediction, computational biology tools, life science automation, BixBench benchmark\n\n## Benchmark: 92.0% on BixBench-Verified-50\n\n

\n \"BixBench\n

\n\n[BixBench](https://github.com/Future-House/BixBench) is a benchmark for evaluating AI agents on real-world bioinformatics tasks. **SciAgent-Skills achieved 92.0% accuracy** on BixBench-Verified-50, the highest among all tested systems:\n\n| System | BixBench-Verified-50 Accuracy |\n|--------|:----------------------------:|\n| Claude Code (Opus 4.6) **+ SciAgent-Skills** | **92.0%** |\n| Claude Code (Opus 4.6) baseline | 65.3% |\n\nSimply equipping Claude Code with these domain-specific skills yields a **+26.7 percentage point improvement** — no fine-tuning, no custom model, just structured scientific knowledge.\n\n## Try It Now — HyperLab\n\nWant to try these skills without any setup? **[HyperLab](https://hyperlab.ai/en/)** (하이퍼랩) is the web platform powered by SciAgent-Skills. Sign up and start analyzing your bioinformatics data directly in your browser — RNA-seq, proteomics, drug screening, and more.\n\n[![Try HyperLab](https://img.shields.io/badge/Try-HyperLab-blue?style=for-the-badge)](https://hyperlab.ai/en/)\n\n---\n\n**199 ready-to-use scientific skills for AI coding agents** — covering genomics, proteomics, drug discovery, biostatistics, scientific computing, and scientific writing.\n\nEach skill is a self-contained SKILL.md file with runnable code examples, key parameters, troubleshooting guides, and best practices. Designed for [Claude Code](https://docs.anthropic.com/en/docs/claude-code), but compatible with any AI agent that reads markdown skill files ([setup guides below](#using-with-other-agents)).\n\n## What's Inside\n\n| Category | Skills | Examples |\n|----------|:------:|----------|\n| Genomics & Bioinformatics | 65 | Scanpy, BioPython, pysam, gget, KEGG, PubMed, scvi-tools, Bakta, Roary |\n| Structural Biology & Drug Discovery | 26 | RDKit, AutoDock Vina, ChEMBL, PDB, DeepChem, datamol |\n| Scientific Computing | 24 | Polars, Dask, NetworkX, SymPy, UMAP, PyG, Zarr, SimPy |\n| Cell Biology | 15 | pydicom, histolab, FlowIO |\n| Biostatistics | 12 | scikit-learn, statsmodels, PyMC, SHAP, survival analysis |\n| Scientific Writing | 21 | Manuscript writing, peer review, LaTeX posters, slides, figure guides |\n| Systems Biology & Multi-omics | 11 | COBRApy, LaminDB, Reactome, STRING |\n| Proteomics & Protein Engineering | 10 | ESM, UniProt, PyOpenMS, matchms, HMDB |\n| Lab Automation | 6 | Opentrons, Benchling |\n| Data Visualization | 5 | Plotly, Seaborn |\n| Molecular Biology | 3 | CRISPR sgRNA design, gene expression, cloning |\n\n**Skill types:** 72 toolkits, 53 database connectors, 37 guides, 37 pipelines\n\n## Installation\n\n### Prerequisites\n\n- An AI coding agent: [Claude Code](https://docs.anthropic.com/en/docs/claude-code), [OpenAI Codex CLI](https://github.com/openai/codex), [Cursor](https://www.cursor.com/), or [Windsurf](https://windsurf.com/)\n- Git\n- Python 3.12+ (only needed if you want to run validation scripts)\n\n> **Note:** SciAgent-Skills is **not** an npm package. Skills are plain markdown files read directly by your AI agent — no `npx`, `npm install`, or runtime dependencies needed. Just clone the repository and point your agent at the skill files.\n\n### Step 1: Clone the Repository\n\n```bash\ngit clone https://github.com/jaechang-hits/SciAgent-Skills.git\ncd SciAgent-Skills\n```\n\n### Step 2: Choose Your Setup Method\n\n#### Method A: Claude Code Plugin (Recommended)\n\nLoad SciAgent-Skills as a Claude Code plugin for the current session:\n\n```bash\nclaude --plugin-dir /path/to/SciAgent-Skills\n```\n\nTo verify the plugin loaded, run `/plugin` inside Claude Code and check that `sciagent-skills` appears in the Installed tab.\n\nSkills become available as `/sciagent-skills:`:\n\n```\n/sciagent-skills:scanpy-scrna-seq\n/sciagent-skills:rdkit-cheminformatics\n/sciagent-skills:pymc-bayesian-modeling\n```\n\nOr just describe your task — the agent finds the relevant skill automatically:\n\n> \"Perform differential expression analysis on this RNA-seq count matrix\"\n\n**Persistent installation** — to load the plugin automatically in every session, use the plugin install command inside Claude Code:\n\n```\n/plugin marketplace add jaechang-hits/SciAgent-Skills\n/plugin install sciagent-skills\n```\n\n#### Method B: Project-Level Integration (Claude Code)\n\nClone into your project directory:\n\n```bash\ncd your-project\ngit clone https://github.com/jaechang-hits/SciAgent-Skills.git .sciagent-skills\n```\n\nAdd the canonical guidance to your project's `AGENTS.md`:\n\n```markdown\n## Scientific Skills\nReference skills in `.sciagent-skills/skills/` for domain-specific analysis.\nRegistry: `.sciagent-skills/registry.yaml`\n```\n\nIf the project does not already have a `CLAUDE.md`, add this compatibility shim so Claude Code imports the same guidance:\n\n```markdown\n@AGENTS.md\n```\n\n#### Using with Other Agents\n\nSciAgent-Skills works with any AI agent that can read project files. Clone the repo into your project, then configure the agent to discover skills via `registry.yaml`.\n\n**Method C: OpenAI Codex CLI**\n\n```bash\ncd your-project\ngit clone https://github.com/jaechang-hits/SciAgent-Skills.git .sciagent-skills\ncp .sciagent-skills/integration-templates/AGENTS.md ./AGENTS.md\n```\n\nCodex reads `AGENTS.md` at the project root automatically. If you already have an `AGENTS.md`, append the contents from the template.\n\n**Method D: Cursor**\n\n```bash\ncd your-project\ngit clone https://github.com/jaechang-hits/SciAgent-Skills.git .sciagent-skills\nmkdir -p .cursor/rules\ncp .sciagent-skills/integration-templates/cursor-rules.md .cursor/rules/sciagent-skills.md\n```\n\nCursor loads rules from `.cursor/rules/` automatically. Alternatively, you can use the `AGENTS.md` template — Cursor supports it as well.\n\n**Method E: Windsurf**\n\n```bash\ncd your-project\ngit clone https://github.com/jaechang-hits/SciAgent-Skills.git .sciagent-skills\nmkdir -p .windsurf/rules\ncp .sciagent-skills/integration-templates/windsurf-rules.md .windsurf/rules/sciagent-skills.md\n```\n\nWindsurf loads rules from `.windsurf/rules/` automatically. Alternatively, you can use the `AGENTS.md` template — Windsurf supports it as well.\n\n**Other agents**: If your agent reads project files, clone the repo as `.sciagent-skills/` and instruct the agent (via its config mechanism) to read `.sciagent-skills/registry.yaml` for skill discovery.\n\n### Step 3: Install Dependencies\n\n```bash\ncd SciAgent-Skills\npixi install\n```\n\n[Pixi](https://pixi.sh) handles the Python environment and all required packages. If you don't have pixi installed:\n\n```bash\ncurl -fsSL https://pixi.sh/install.sh | bash\n```\n\n## How Skills Work\n\nEach skill follows a structured template:\n\n```\nskills///\n SKILL.md # Main skill file (300-550 lines)\n references/ # Optional deep-dive reference files\n assets/ # Optional templates, configs\n```\n\nA **SKILL.md** contains:\n\n- **Frontmatter** — name, description, license (for agent discovery)\n- **Overview & When to Use** — what the tool does and when to reach for it\n- **Prerequisites** — packages, data, environment setup\n- **Quick Start** — minimal copy-paste example\n- **Workflow / Core API** — step-by-step pipeline or module-by-module API guide\n- **Key Parameters** — tunable settings with defaults and ranges\n- **Common Recipes** — self-contained snippets for common tasks\n- **Troubleshooting** — problem/cause/solution table\n\nThe agent reads only the `description` field during planning. Full skill content is loaded on demand when relevant.\n\n## Directory Structure\n\n```\nSciAgent-Skills/\n├── .claude-plugin/\n│ └── plugin.json # Claude Code plugin manifest\n├── integration-templates/ # Config templates for Codex, Cursor, Windsurf\n├── skills/ # All 199 skills organized by category\n│ ├── genomics-bioinformatics/\n│ ├── structural-biology-drug-discovery/\n│ ├── scientific-computing/\n│ ├── cell-biology/\n│ ├── biostatistics/\n│ ├── scientific-writing/\n│ ├── systems-biology-multiomics/\n│ ├── proteomics-protein-engineering/\n│ ├── lab-automation/\n│ ├── data-visualization/\n│ └── molecular-biology/\n├── templates/ # Skill authoring templates\n├── registry.yaml # Index of all skills\n├── AGENTS.md # Skill authoring guide\n└── scripts/\n └── validate_registry.py\n```\n\n## Example Use Cases\n\n**Single-Cell RNA-seq Analysis** (scRNA-seq)\n> \"Load 10X data, QC filter, normalize, cluster, find marker genes, and annotate cell types\"\n\nUses: `anndata-data-structure` → `scanpy-scrna-seq`\n\n**Bulk RNA-seq Differential Expression**\n> \"Run DESeq2-style differential expression analysis on this count matrix, generate volcano plot\"\n\nUses: `pydeseq2-differential-expression` → `matplotlib-scientific-plotting`\n\n**Drug Discovery Pipeline**\n> \"Search ChEMBL for EGFR inhibitors with IC50 < 100nM, filter with Lipinski rules using RDKit, dock top candidates with AutoDock Vina\"\n\nUses: `chembl-database-bioactivity` → `rdkit-cheminformatics` → `autodock-vina-docking`\n\n**Protein Structure Prediction & Analysis**\n> \"Get the AlphaFold structure for UniProt P04637, assess confidence, find high-confidence binding regions\"\n\nUses: `uniprot-protein-database` → `alphafold-database-access`\n\n**Bayesian Biostatistics**\n> \"Fit a hierarchical Bayesian model to this clinical trial data with patient-level random effects\"\n\nUses: `pymc-bayesian-modeling` → `matplotlib-scientific-plotting`\n\n**Multi-omics Integration**\n> \"Integrate transcriptomics and proteomics data, run pathway enrichment, build a protein interaction network\"\n\nUses: `lamindb-data-management` → `reactome-pathway-analysis` → `string-protein-interaction`\n\n## Contributing\n\n### Adding a New Skill\n\n1. Read `AGENTS.md` for the full authoring workflow\n2. Classify your topic (pipeline / toolkit / database / guide)\n3. Pick a category from the table above\n4. Use the appropriate template from `templates/`\n5. Add the entry to `registry.yaml`\n6. Validate: `python scripts/validate_registry.py`\n\n### Skill Templates\n\n| Template | Use When |\n|----------|----------|\n| `SKILL_TEMPLATE.md` | Linear input→process→output pipeline (e.g., DESeq2) |\n| `SKILL_TEMPLATE_TOOLKIT.md` | Collection of independent modules (e.g., RDKit) |\n| `SKILL_TEMPLATE_PROSE.md` | Conceptual guide, decision frameworks (e.g., statistical test selection) |\n\n## Requirements\n\n- Python 3.12+ (for validation scripts)\n- No runtime dependencies — skills are markdown files read by the agent\n- Individual skills list their own tool-specific prerequisites (e.g., `pip install scanpy`)\n\n## Comparison with Other Tools\n\n| Feature | SciAgent-Skills | GPTomics/bioSkills | ClawBio |\n|---------|:-:|:-:|:-:|\n| Total skills | **199** | ~30 | ~20 |\n| BixBench benchmark | **92.0%** | — | — |\n| Skill types | Pipeline, Toolkit, Database, Guide | Pipeline | Pipeline |\n| Multi-agent support | Claude Code, Codex, Cursor, Windsurf | Claude Code | Claude Code |\n| Claude Code plugin | Yes | No | No |\n| Web platform | [HyperLab](https://hyperlab.ai/en/) | No | No |\n\n## License\n\nCC-BY-4.0 for original content. Individual skills note the license of their underlying tools.\n\n## Acknowledgments\n\nThis project builds on 50+ open-source scientific Python packages. If you find a skill useful, consider starring the underlying tool's repository and supporting its maintainers.\n\n---\n\n**Related searches:** bioinformatics AI agent, Claude Code scientific skills, RNA-seq analysis tool, single-cell RNA-seq AI, drug discovery AI pipeline, protein structure prediction, computational biology automation, life science AI tools, 바이오인포매틱스 AI, 오믹스 호라이즌, 생명과학 AI 에이전트, BixBench benchmark\n", "bytes": 13093, "sha": "db285b587002175f384852cd55b8e43f89036401f9b704de2ba24958700432a2", "repo_slug": "jaechang-hits/sciagent-skills", "fonte": "repo", "truncated": false, "api": "https://agentalog.com/api/listings/plg_jaechang_hits_sciagent_skills_sciagent_s_b994dc7a/readme" }