# Manage Biological Data with LaminDB

Biological data workflows often lose metadata, lineage, and consistent annotations. This skill provides LaminDB guidance for organizing, validating, querying, tracing, integrating, and deploying research data.

## Install

```bash
npx skillstore add davila7/lamindb
```

## Metadata

- Status: approved
- Slug: davila7-lamindb
- Skillstore revision: r2
- Version status: missing
- Tree hash: ac584aac6d549cac6e40a91438541b86d253552fd370f53c3a8fcf36a6831ee9
- Author: davila7
- GitHub username: davila7
- License: MIT
- Repository: https://github.com/davila7/claude-code-templates/tree/main/cli-tool/components/skills/scientific/lamindb
- Ref: c43861a65bb95efcae259cd161c9d6f4dc7eec6f
- Supported tools: Claude, Codex, Claude Code
- Audit status: complete
- Agent install advisory: allowed
- Manual install advisory: allowed
- Artifact signature: available
- Audit attestation: unavailable
- Human verification: not\_verified
- Risk factors: filesystem, network, env\_access, external\_commands
- Quality score: 50
- Quality tier: warning
- Public page: https://skillstore.pages.dev/skills/davila7-lamindb
- Manifest: https://skillstore.pages.dev/api/skills/davila7-lamindb/manifest

## Capabilities

- Explains how to create, version, query, stream, and organize LaminDB artifacts and collections.
- Guides schema validation for DataFrame, AnnData, MuData, SpatialData, and TileDB-SOMA datasets.
- Shows ontology workflows for genes, cell types, tissues, diseases, and other Bionty registries.
- Provides lineage patterns for notebooks, scripts, Nextflow, Snakemake, and Redun workflows.
- Documents integrations with W&B, MLflow, Hugging Face, scVI-tools, Vitessce, DuckDB, and Git.
- Covers local, SQLite, PostgreSQL, S3, GCS, MinIO, and Cloudflare R2 setup options.

## Use Cases

- Curate single-cell datasets: Build validation and ontology workflows for AnnData metadata before analysis or sharing.
- Build a biological data lakehouse: Organize versioned artifacts, searchable features, collections, and lineage across local or cloud storage.
- Track computational pipelines: Connect workflow runs, inputs, outputs, parameters, models, and experiment platforms for reproducible research.

## Prompt Templates

### Create a local instance

```
Help me install LaminDB, initialize local storage at [path], register [dataset type], and explain each verification step.
```

### Curate an annotated dataset

```
Design a LaminDB curation workflow for [dataset type] with required columns [columns] and ontology-backed fields [fields].
```

### Track a research pipeline

```
Plan a tracked [workflow manager] pipeline that loads [inputs], records parameters, saves [outputs], and preserves code and data lineage.
```

### Design production infrastructure

```
Design a production LaminDB deployment using [storage], [database], [identity provider], and [regions], including migration, access control, backups, and monitoring.
```

## Limitations

- Provides guidance and examples but does not execute LaminDB operations by itself.
- Requires LaminDB and access to each configured database, storage system, and third-party service.
- Examples may require adaptation for your LaminDB version, schema, permissions, and infrastructure.
- Does not replace security, privacy, or regulatory review for biological or clinical data.

## Best Practices

- Define schemas and ontology-backed features before registering large datasets.
- Track every workflow run with explicit inputs, parameters, outputs, and descriptive artifact metadata.
- Query metadata before loading files, and use private encrypted storage for sensitive research data.

## Anti Patterns

- Do not hardcode credentials, default passwords, or database secrets in commands and notebooks.
- Do not delete caches, artifacts, or instances without validating scope and preserving required backups.
- Do not treat example deployment settings as production-ready without reviewing permissions, encryption, privacy, and compliance.

## Security Audit

- Audited at: 2026-07-23T12:45:55.753\+00:00
- Summary: Most analyzer hits are false positives caused by the Python alias ln, Markdown backticks, placeholders, and documentation links. Confirmed risks include default MinIO credentials, plaintext HTTP, permissive shared-cache access, shared temporary storage, and unguarded recursive deletion. Database passwords in command arguments and unprotected database backups are additional context-level risks.

## Stats

- Views: 519
- Downloads: 8
- Favorites: 0
- Popularity score: 0