Guide · Agent skill

How to use the Biopython agent skill: setup, examples and tips

Updated 2026-10-29 · written with AI from the project's own documentation and checked against our catalogue data

The Biopython agent skill gives an AI agent working knowledge of Biopython, a widely used Python library for biological computation. Once it is installed, you can ask the agent in plain language to read or convert sequence files, search NCBI databases, run BLAST, or analyse protein structures, and it will write and run code that follows the skill's reference guides. This guide is for readers who have never used an agent skill, and it covers installation, triggering, examples, requirements, safety and tips.

Biopython Molecular Biology ToolkitK-Dense-AI/scientific-agent-skills · MIT Download the skill

Quick start

1Download the ZIP. Download the Biopython skill ZIP from this page.
2Unzip into a skills folder. Unzip it into ~/.claude/skills/ for personal use, or into a project's .claude/skills/ folder for one project.
3Or upload in the Claude apps. If you use the Claude apps, upload the ZIP under Settings → Capabilities → Skills.
4Install Biopython. Install the pinned Biopython release in your Python environment with the uv command shown above.
5Set your NCBI email. Set NCBI_EMAIL, or set Entrez.email in your script, before any NCBI database call.
6Ask for a task. Describe the file or database, the operation and the output you want in plain language.

What does the Biopython skill make the agent do?

An agent skill is a folder of instructions and reference files that an AI agent loads when a task matches it. The Biopython skill tells the agent how to use Biopython's sub-packages for sequence handling, alignment, database access through Bio.Entrez, BLAST, structure analysis with Bio.PDB, and phylogenetics with Bio.Phylo.

The skill covers reading and writing FASTA, GenBank, FASTQ, PDB and mmCIF files, sequence statistics such as GC content and molecular weight, and motif analysis. It also includes guidance for batch processing and custom pipelines.

Read, convert and extract sequences from FASTA, GenBank, FASTQ, PDB and mmCIF files
Run pairwise alignments and read or write multiple sequence alignments
Search and download records from PubMed, GenBank, Protein and Gene through Bio.Entrez
Run BLAST searches via NCBI web services or parse BLAST XML output
Analyse protein structures, calculate distances and angles, and superimpose structures
Build, read and manipulate phylogenetic trees

How do you install the Biopython skill?

Download the skill ZIP from this page. Then choose one of two routes: unzip it into your personal skills folder at ~/.claude/skills/, or unzip it into a project's .claude/skills/ folder if you want the skill available only in that project.

If you use the Claude apps instead, upload the ZIP under Settings → Capabilities → Skills.

The skill itself also expects the Biopython library to be installed in your Python environment. The skill's own installation guidance pins a specific Biopython release for reproducibility, using this command:

uv pip install "biopython==1.88"

How do you set up NCBI email and API key?

For any NCBI database access, set your real email address, because NCBI requires it for Entrez calls. The skill reads the NCBI_EMAIL environment variable and treats it as recommended rather than strictly required in the metadata, but you should set it for any Entrez work.

For higher rate limits, the skill uses NCBI_API_KEY. With a key, Entrez allows 10 requests per second instead of 3. The skill's guidance is to read only NCBI_API_KEY from the environment and never hardcode keys or load unrelated environment variables.

NCBI_EMAIL: your contact email, recommended and used for Entrez identification
NCBI_API_KEY: optional key that raises Entrez rate limits from 3 to 10 requests per second

import os
from Bio import Entrez

Entrez.email = "your.email@example.com"  # required — use your real email
Entrez.tool = "your_tool_name"  # optional but recommended for reusable software

if api_key := os.environ.get("NCBI_API_KEY"):
    Entrez.api_key = api_key

How do you trigger the skill with a natural request?

You do not need special syntax. Describe the biological task and the data involved, and the agent will decide whether the skill applies, based on the task areas it lists. Be specific about file types, the database you want, and where results should go.

Good requests name the input, the operation and the output. For example, ask for a script that reads a named FASTA file and writes a summary CSV, rather than a general request to 'analyse my sequences'.

'Read sequences.fasta and print each record ID with its length in bp.'
'Convert input.gb to FASTA and save it as output.fasta.'
'Search PubMed for recent biopython papers and list titles and PMIDs in a table.'
'Parse the BLAST XML output and keep hits with E-value below 1e-5.'

What requirements and safety notes should you know?

The skill requires Python 3.10 or newer, NumPy and Biopython. Plotting is optional and needs Matplotlib or ReportLab. Entrez and web BLAST need network access and a contact email. Local BLAST, MUSCLE 5, Clustal Omega and DSSP each need separate executables; the agent does not install them.

The skill runs scripts and uses the network. Remote NCBI services have rate limits, and the skill warns that BLAST has its own pacing rules, including leaving at least 10 seconds between contacts. The examples are illustrative and are not validated for clinical or diagnostic use.

Python 3.10 or newer, NumPy and Biopython are required
Optional: Matplotlib or ReportLab for plots
Optional local tools: BLAST, MUSCLE 5, Clustal Omega, DSSP
Runs scripts: yes. Uses network: yes
Do not use outputs for clinical or diagnostic decisions

What tips help you get reliable results?

Ask the agent to check its code against the example files and reference guides the skill ships with. The skill's review notes record which examples were exercised locally and which depend on your own files or external programs, so treat unlisted examples as starting points.

For large NCBI or BLAST jobs, ask for batching and delays between requests, and set your email and tool name first. Keep sequence files and output folders in your project so the script can be rerun.

Pin your Biopython version so results stay reproducible
Set NCBI_EMAIL before any Entrez call
Start with small batches before large downloads
Check which external executables you have before asking for local BLAST or alignment tools

What people use it for

Converting a folder of GenBank files to FASTA for a downstream tool
Downloading PubMed or GenBank records for a list of search terms, with rate limits respected
Writing a batch script that summarises sequence length and GC content across FASTA files
Parsing BLAST XML output and filtering hits by E-value or identity
Extracting sequences and computing distances from a PDB structure

Frequently asked questions

Is the Biopython agent skill free?

This listing states an MIT licence for the skill. The skill file itself names the Biopython License Agreement, so check that file for the terms that apply to the bundled material. Biopython is described in the source as freely available.

What do I need to run the Biopython skill?

You need Python 3.10 or newer, NumPy and Biopython. Entrez and web BLAST need network access and a contact email. Local BLAST, MUSCLE 5, Clustal Omega and DSSP need their own installs if you want to use them.

How do I install an agent skill like this one?

Download the ZIP and unzip it into ~/.claude/skills/ or a project's .claude/skills/ folder. In the Claude apps, you can upload the ZIP under Settings → Capabilities → Skills instead.

Why does the agent need my email address?

NCBI requires a contact email for Entrez calls, so the skill uses NCBI_EMAIL or Entrez.email to identify your requests. Setting a real address is recommended for any NCBI database access.

How does this compare with using Biopython directly?

The skill does not replace Biopython. It gives the agent reference guides and examples so it can write Biopython code for you. If you prefer to write and run code yourself, you can use the Biopython documentation directly.

Why did my NCBI or BLAST request fail or get slow?

Remote services have rate limits. Set NCBI_API_KEY to raise Entrez limits from 3 to 10 requests per second, and leave pauses between BLAST requests as the skill advises. Also confirm your email is set.

New agent skills, repos and AI tools — every week

The best new skills and open-source projects, the fastest-growing repos, and one free prompt. Five minutes to read. Unsubscribe anytime.

Our verdict

Use this skill if you are learning to script sequence analysis, want an agent to write batch scripts for FASTA, GenBank or NCBI downloads, or are building Python molecular biology pipelines. Pick something else if you need clinical or diagnostic results, or if you expect the agent to run large NCBI or BLAST jobs without attention to rate limits and external tool installs.

Download the skill