Skip to content

Guide · Agent skill

How to use the Primer Design and Specificity agent skill

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

The primer-design skill lets an AI agent design PCR and RT-qPCR primer pairs from a DNA sequence you supply, check primer pairs you already have, and screen candidate primers against reference sequences for likely off-target amplification. This guide explains what the skill does, how to install it, how to ask for work in plain language, and what it needs. After reading it you should be able to run a first design and understand what the results do and do not prove.

Primer Design and SpecificityK-Dense-AI/scientific-agent-skills · MIT

Quick start

  1. 1Download and install the skill. Download the ZIP from this page and unzip it into ~/.claude/skills/ or a project's .claude/skills/ folder, or upload it under Settings, then Capabilities, then Skills.
  2. 2Set up Python and primer3-py. Create a virtual environment with Python 3.11 or newer and install assets/requirements.txt into it. Run design_primers.py --help to confirm it loads.
  3. 3Prepare your reviewed target sequence. Gather the target FASTA, its accession with version, the product size you need and any regions to avoid. Note the source and checksum for your report.
  4. 4Run the design. Ask the agent to design primers, or run design_primers.py with your template, a preset and the qPCR config. Check the outputs design.json, pairs.tsv and expected.tsv.
  5. 5Screen for specificity. Screen candidates against the relevant reference sequences, optionally with BLAST+ installed. Confirm each expected product maps to the right coordinates.
  6. 6Record the evidence. Fill in assets/assay-report-template.md with the reference manifest, reaction assumptions and settings, and label each pair as a computational candidate.

What the primer-design skill makes the agent do

Agent skills are folders of instructions and scripts that an agent loads when a request matches them. This one covers primer work: designing new primer pairs, auditing existing ones, and screening primers for specificity. Melting temperatures and secondary structure are calculated with the Primer3 library under the salt and oligo conditions you state.

The skill keeps coordinates, settings and engine versions in a traceable report, and it labels every result as a computational candidate rather than an experimentally validated assay. A good Primer3 score or a single BLAST hit is not treated as proof of specificity.

  • Design PCR and RT-qPCR primer pairs from a supplied sequence
  • Check both full oligos and annealing cores of existing pairs
  • Screen candidates against reference sequences with bounded local in-silico PCR, and optionally BLAST
  • Handle exon-junction or isoform-specific assays, variant masking, cloning tails and multiplex compatibility
  • Interpret Primer-BLAST results

Install the primer-design skill

Download the ZIP for this skill from this page. Then either unzip it into your personal skills folder, ~/.claude/skills/, or into a project's .claude/skills/ folder so it applies only to that project.

If you use the Claude apps, you can instead upload the skill under Settings, then Capabilities, then Skills.

After installing, set SKILL_DIR to the folder where the skill is actually installed. Keep analysis work in a separate directory so that environments, reference files and results do not end up inside the skill.

  • Personal: ~/.claude/skills/
  • Project only: .claude/skills/ inside your project
  • Claude apps: Settings, then Capabilities, then Skills, then upload

Set up the Python environment and check it

The design and thermodynamics scripts need Python 3.11 or newer and the primer3-py package pinned in assets/requirements.txt. The source example creates a virtual environment with uv and installs that file into it. Run the help command to confirm the scripts load before you rely on them.

On Windows, use the environment's Scripts/python.exe instead of bin/python.

  • Python 3.11 or newer is required
  • Pinned dependency: primer3-py, listed in assets/requirements.txt
  • The local exhaustive screen uses only the standard library
uv venv --python 3.13 .venv-primer uv pip install --python .venv-primer/bin/python -r "$SKILL_DIR/assets/requirements.txt" .venv-primer/bin/python "$SKILL_DIR/scripts/design_primers.py" --help

Run your first primer design

For a real project, give the agent or script your reviewed target FASTA and your constraints. If the FASTA has several records, you must name the exact record ID with --record. The pcr preset requests 100 to 1000 bp products and the qpcr preset requests 70 to 200 bp; you can override these ranges in the configuration file.

The tool passes validated settings to Primer3, rejects misspelled settings rather than ignoring them, and checks returned sequences against the template. If no candidates come back, read the engine explanations and change a biologically justified constraint instead of relaxing everything at once.

  • Use --record when the FASTA contains multiple sequences
  • Use --mask-bed only for exclusions in the template's own coordinates
  • Use --forward-tail and --reverse-tail for cloning tails; tails are checked in the complete-oligo step, not in the design score
  • Keep earlier reports when you compare parameter choices
.venv-primer/bin/python "$SKILL_DIR/scripts/design_primers.py" --template your_target.fasta --preset qpcr --config "$SKILL_DIR/assets/qpcr-config.json" --output design.json --pairs-out pairs.tsv --expected-out expected.tsv

Screen primers for off-target amplification

Designing a primer pair does not show it amplifies only your target. Screen candidates against the reference sequences that matter for your sample, such as paralogs, pseudogenes, alternate contigs, transcript isoforms, vector backbone or host sequence. Record each reference's accession with version and its checksum in the evidence report.

The optional BLAST screen needs NCBI BLAST+ installed from its official distribution, with blastn and makeblastdb available on your PATH. You can confirm both with the commands below. Network access is needed only to retrieve references or to use external Primer-BLAST, not to run the local scripts.

  • Use the supplied assets/assay-report-template.md to collect evidence
  • Map expected product coordinates to each screening reference's record ID and orientation
  • Treat BLAST+ output as one piece of evidence, checked against the stated search limits
blastn -version makeblastdb -version

Know what the skill does not cover

The bundled tools handle ordinary DNA primers with unambiguous ACGT cores. Degenerate, bisulfite, probe and modified-base assays need specialized review that the bundled ordinary-DNA model does not provide. Guide-RNA design and expression normalization are excluded from this skill.

The results are computational. Any diagnostic or clinical use needs experimental validation that the skill cannot perform.

  • Not supported: degenerate, bisulfite, probe or modified-base assays
  • Excluded: guide-RNA design and expression normalization
  • Results are in-silico candidates until tested in the lab

What people use it for

  • Designing a qPCR primer pair for a gene of interest from a reviewed reference sequence, with the product size range stated in the report
  • Auditing primer pairs already used in a lab against a new reference genome or a new set of transcript isoforms
  • Designing primers that span an exon junction or distinguish one isoform, after supplying the annotation that defines the junction
  • Adding cloning tails to annealing cores and checking the complete oligos and the final product
  • Checking multiplex panels for oligo interactions and cross-pair products

Frequently asked questions

Is the primer-design agent skill free to use?

The skill is published under the MIT licence, as stated on this page. Check the licence file included in the download for the full terms before you reuse or redistribute it.

What do I need installed to run the primer-design skill?

You need Python 3.11 or newer and the primer3-py package from assets/requirements.txt. The optional BLAST screen also needs NCBI BLAST+ with blastn and makeblastdb on your PATH. Network access is only needed to retrieve references or use external Primer-BLAST.

How do I install an agent skill from a ZIP file?

Unzip the downloaded file into ~/.claude/skills/ for personal use, or into a project's .claude/skills/ folder. In the Claude apps you can instead upload it under Settings, then Capabilities, then Skills.

Can the skill design primers for qPCR?

Yes. The qpcr preset requests products of 70 to 200 bp by default, and you can override that range in the configuration. The skill also supports RT-qPCR, which is described in its design workflows reference.

Does the skill guarantee that my primers are specific?

No. It checks specificity against the reference sequences you supply and reports computational candidates. The skill states that a good Primer3 score or a single BLAST alignment cannot establish specificity, so experimental validation is still needed.

Does the skill run anything over the internet?

The local design and screening scripts use files on your machine and do not submit sequences online. The skill lists network access only for installation, reference retrieval or external Primer-BLAST.

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 preparing a PCR or RT-qPCR assay and want a documented, reproducible design record, or if you are learning how Primer3 settings change results. Pick another tool if you need degenerate, bisulfite, probe or modified-base primers, guide-RNA design, or results you can use diagnostically without wet-lab validation.

Download the skill

Find exactly what
you're looking for