Skills
gget skill review: lightweight genomic queries for Claude
gget is a mature bioinformatics CLI (185k GitHub stars) packaged as a Claude Code skill with zero marketplace adoption and no ratings. Useful for quick gene lookups and sequence searches, but lacks documented agent workflows and has seen zero real-world Claude integration testing.
On the agentskill.sh marketplace, a skill is a structured, agent-aware guide paired with version metadata and auto-feedback hooks. Unlike a generic README, a skill teaches an AI agent (in this case, Claude Code) when to invoke a tool, what failure modes to expect, and how to log results for reproducibility. The gget skill (owned by affaan-m) wraps the gget bioinformatics package as a Claude Code capability.
Marketplace Signals at a Glance
| Metric | Value | Interpretation |
|---|---|---|
| Agentskill rating | 0 out of 5 | No agent feedback submitted yet |
| Agentskill installs | 0 | Not adopted by Claude Code users |
| Security score | 97/100 | Upstream package passes automated checks |
| Content quality | 92/100 | SKILL.md is detailed and well-structured |
| Upstream GitHub stars | 185,629 | gget is a mature, widely-known bioinformatics project |
| Last updated | 2026-05-21 | Recently indexed; may be newly published |
The glaring mismatch: 185k upstream stars but zero marketplace installs and zero ratings. This signals either a newly published skill, poor discoverability, or a mismatch between who uses gget (bioinformaticians, academics) and who uses Claude Code agents (developers, ops teams, product builders). The high security and content scores confirm the SKILL.md itself is solid; the friction is adoption and validation.
What gget Actually Does
gget is a Python CLI and library that provides “quick genomic database queries” across Ensembl, UniProt, NCBI BLAST, protein structure (AlphaFold, PDB), and specialized databases like OpenTargets, COSMIC, and Bgee. The package bundles around a dozen modules: search (find genes), info (metadata), seq (sequences), ref (reference genome links), blast (rapid alignment), and domain-specific queries like enrichr (enrichment), cbio (cBioPortal), and alphafold (protein structure prediction).
The SKILL.md describes the use case tightly: “quick” is the operative word. gget is not a replacement for a full bioinformatics pipeline. It is a first-pass lookup tool: verify a gene ID, grab a sequence, run a small BLAST alignment, or fetch reference links before committing to slower, more exhaustive tools like Snakemake, Nextflow, or BLAST+.
Installation and Integration
On Claude Code, after installing the skill from agentskill.sh, it would be accessible via .claude/skills/affaan-m/scientific-pkg-gget/SKILL.md. The SKILL.md itself provides clear installation steps: create a virtual environment, upgrade pip, and run pip install gget. The skill also notes uv as an alternative, which signals some attention to modern Python packaging practices.
No supporting files are mentioned in the SKILL.md. The skill is self-contained: it is a guide, not a bundle of scripts or pre-downloaded data. When Claude Code invokes gget, it calls the live CLI or imports the Python module directly, making real-time queries to upstream databases.
Platform compatibility is listed as “openclaw” (presumably the Claude Code backend). The skill is marked as an API-type skill, meaning it expects network calls to external services. This is reasonable for a bioinformatics lookup tool; gget has no offline mode for large databases.
Capabilities and Scope
The SKILL.md defines three explicit contexts for when to use gget:
- Finding Ensembl IDs, gene metadata, transcripts, or sequences.
- Running quick BLAST or BLAT lookups without local infrastructure.
- Querying reference genome links, pathways, cancer associations, or expression data through a single interface.
A fourth strength is the “reproducible evidence log” framing: the SKILL.md includes a template for documenting module, version, species, assembly, and output format. This is appropriate for scientific workflows where reproducibility is non-negotiable.
The skill also makes clear what not to use gget for: regulated clinical interpretation, production-scale throughput, or fine-grained control over database versions. These guardrails are mature and honest.
Comparison with Alternatives
For Claude Code agents needing bioinformatics lookups, three paths exist:
| Approach | Use Case | Strengths | Weaknesses |
|---|---|---|---|
| gget (this skill) | Quick Ensembl, BLAST, protein lookups | Unified CLI, live API, lightweight | No offline data, rate-limit dependent, niche adoption |
| Biopython (default Claude capability) | Sequence parsing, local alignment, custom workflows | Full Python library, fine-grained control | Requires local sequences, steeper learning curve, slower |
| Direct API calls (no skill) | Custom database queries via REST | Full control, minimal abstraction | No structured error handling, reinvent metadata tracking |
gget sits between Biopython (too heavy for a quick query) and raw API calls (no structure). For an agent that needs to ask “what is the BRCA1 gene ID?” or “fetch the human reference genome version GRCh38,” gget is ideal. For anything else, Biopython or specialized tools are more appropriate.
Gaps and Risks
Several friction points emerge on close reading of the SKILL.md and the marketplace metrics:
Zero agent integration examples. The SKILL.md shows CLI and Python snippets but does not demonstrate how a Claude Code agent should trigger or parse gget output. In a multi-step agent workflow (e.g., “search for a gene, fetch its sequence, run BLAST on the sequence”), the skill does not document how to chain these steps or handle partial failures. This is a significant omission for an agent-focused marketplace.
No documented failure modes. The SKILL.md warns to upgrade gget but does not explain what happens when a database is down, a gene ID is invalid, or a BLAST query times out. Agents need explicit error detection and recovery logic; the skill provides a checklist but not error handling patterns.
Dependency narrowness. The SKILL.md notes “Do not assume every module supports every Python version or dependency set.” This is a red flag. gget likely has pinned scientific dependencies (numpy, pandas, requests) with narrow version compatibility. A Claude Code agent invoking gget across many runs may hit unresolved dependency conflicts. The skill does not provide a locked requirements file or dependency resolution strategy.
Untested at scale. Zero installs mean zero feedback. No one has actually run this skill in Claude Code and reported success or failure. The upstream gget project is solid (185k stars, published, cited), but the agent workflow integration is unproven.
Weak SEO and discoverability. The job categories are listed as “education-research, data-analytics, development.” That is broad and vague. A Claude Code user searching for “sequence alignment” or “gene lookup” might not discover gget if the skill is tagged only with these generic terms. Tighter categorization (e.g., “genomics-tools”, “ensembl-integration”, “bioinformatics-agents”) would help.
When to Use This Skill
Use gget via this skill if:
- You are prototyping a bioinformatics research agent and need to sketch out gene lookups before building a real pipeline.
- You need to verify a gene ID, fetch a reference genome link, or run a quick BLAST alignment as part of a larger Claude Code workflow.
- You are comfortable with Python virtual environments and live API queries.
- Your task does not require offline operation, clinical validation, or batch processing at scale.
Skip this skill if:
- Your query involves regulated or clinical data (use a specialized clinical genetics tool or consult a human expert).
- You need offline operation or full control over database versions (use Snakemake, Nextflow, or local BLAST+).
- Your team is unfamiliar with bioinformatics terminology and database schemas (gget assumes domain knowledge).
- You need real-time support or a large installed user base (zero installs means zero community momentum).
Takeaways
The gget skill wraps a mature, well-known bioinformatics tool but shows zero adoption in the Claude Code marketplace. The SKILL.md is well-written and honest about scope (it emphasizes “quick” lookup, not production pipelines), but it lacks agent-specific integration patterns, error-handling guidance, and documented failure modes. The 185k upstream GitHub stars demonstrate that gget is a legitimate project; the zero installs and zero ratings suggest that either the skill was just published, or there is a mismatch between the bioinformatics audience and the Claude Code agent user base. Security and content quality scores are strong, but real-world agent workflows remain untested. Before adopting this skill in production, seek explicit documentation of how a Claude Code agent should chain gget calls, handle timeouts, and log results reproducibly.
Further reading
- gget GitHub repository: The upstream project; 185k stars, active maintenance, full module documentation.
- gget documentation: Official docs with API reference and examples for each module.
- gget Bioinformatics paper: Peer-reviewed publication describing the tool’s design and use cases.
- agentskill.sh listing for gget: The skill page; submit feedback or request features here.
- Ensembl reference: Background on the main genomic database queried by gget.
Frequently asked
What is a skill on agentskill.sh and how does it differ from just installing gget via pip?
A skill is a structured knowledge file (.claude/skills/...) that teaches Claude Code agents when and how to use a tool, paired with auto-feedback integration. gget as a skill adds reproducibility metadata, environment isolation, and agent-triggered feedback, but the underlying gget CLI is unchanged. The skill is a workflow wrapper, not a reimplementation.
Can I use gget via this skill on any language or just Python?
gget works natively as a CLI (bash/shell) and as a Python package. The SKILL.md covers both patterns and recommends Python virtual environments. Claude Code can invoke either, but the skill assumes Python-first for reproducible logs and version tracking.
Why zero installs and zero ratings after being listed for months?
The skill was indexed at 2026-05-21. Zero installs likely means either the skill was just published, no one in the Claude Code user base has queried genomic data yet, or the listing is not discoverable (poor search metadata, unclear job categories, or niche appeal). The 97/100 security score and 92/100 content quality suggest the SKILL.md is well-written; adoption is the missing variable.
Should I use gget for clinical or production bioinformatics pipelines?
No. The SKILL.md explicitly warns against using gget for regulated clinical interpretation or high-throughput production pipelines. Use dedicated tools like Snakemake, Nextflow, or BLAST+ for those. gget is a quick reference tool, best for hypothesis generation and evidence collection before deeper analysis.
How current are the genomic databases that gget queries?
gget queries live upstream databases (Ensembl, UniProt, NCBI BLAST) via API. Database freshness depends on the vendor's refresh cycle. The SKILL.md warns that databases change over time and recommends upgrading gget regularly and checking upstream docs before each major run to avoid stale or shifted identifiers.