A small, self-contained set of Claude skills for working with an electronic lab notebook (Labstep) and routine nucleic-acid QC data. These are generic, reusable building blocks — point them at your own Labstep workspace, data directory, and templates.
| Skill | What it does |
|---|---|
labstep |
Read and manage Labstep experiments, protocols, resources, and inventory via the lab-mcp client (CLI, MCP server, or Python library). |
nucleic-acid-analysis |
Read and extract nucleic-acid quantification metrics (Qubit, TapeStation, qPCR) from per-experiment data folders, with fuzzy column matching and robust value coercion. |
experiment-summary |
Generate a filled-in Experiment Summary .docx from Labstep metadata and QC files, with full data provenance. |
lab-note |
Turn voice notes, photos, and data files delivered from a chat platform (Telegram/Slack) into context-aware comments on the matching Labstep experiment. |
Drop the skills/ directory into your agent's skills path (e.g. .claude/skills/), or copy individual skills. Each SKILL.md is self-describing.
Most skills lean on the lab-mcp Labstep client:
pip install "lab-mcp @ git+https://github.com/neurogenomics/lab-mcp.git"Per-skill Python dependencies are listed in each SKILL.md.
Copy .env.example to .env and fill in your own values:
cp .env.example .env| Variable | Used by | Purpose |
|---|---|---|
LABSTEP_API_KEY |
all | Labstep API key |
FIREWORKS_API_KEY |
lab-note |
Transcription API key |
DATA_DIR |
nucleic-acid-analysis, experiment-summary |
Root directory holding per-experiment data folders |
EXP_PREFIX |
lab-note |
Your lab's experiment custom-identifier prefix (default EXP) |
The example experiment identifiers (EXP123), column names, and template structure in these skills are illustrative — adapt them to your own assays, naming conventions, and Word templates.
- Data directories are treated as strictly read-only. Skills never write, move, rename, or delete source files; all output goes to the working directory.
- Labstep writes (creating experiments, posting comments) should be confirmed with the user unless you explicitly opt an automated pipeline into unattended writes.
MIT — see LICENSE.