Ready to supercharge your AI agents?

Install any skill in seconds. Free to start, no credit card required.

Get Started Free →

🦖 VCF Annotator · FreedomIntelligence | DecimalAI

Vcf Annotator

NewSkillScoreMaintained

vcf-annotator

Annotate VCF variants with VEP, ClinVar, gnomAD frequencies, and ancestry-aware context. Generates prioritised variant reports.

scienceby FreedomIntelligence

Links this skill to your agents and keeps it current with the author’s updates.

Automatic

Research pack

1 · One-time — install the CLI (docs ↗)

pip install decimalai

2 · Get the skill — writes .claude/skills/vcf-annotator/SKILL.md

decimalai skills pull vcf-annotator --out .claude/skills/

No CLI? Get the same file — curl · raw file

curl -fsSL --create-dirs -o .claude/skills/vcf-annotator/SKILL.md https://app.decimal.ai/s/vcf-annotator/SKILL.md

Raw file: /s/vcf-annotator/SKILL.md ↗ · machine-readable summary at /s/vcf-annotator.json ↗

In productionGlobal

how this skill behaves on real traffic · last 30 days

No production data for this skill yet — usage numbers appear here once agents start running it.

Anonymized aggregate across every org using this skill — proof from real production use, not a demo.

Does this skill help? · Benchmark: with vs. withoutsign in to run this on your agents

22 test cases from the publisher’s eval.yaml · each run twice on gemini-3.6-flash — with the skill injected, then without · last run 59d ago

No measurable change-2% tokens0% turns9% of whole cases pass with skill (2/22)

Test case	Without → With	Effect	Δ tokens	Δ turns
case-10	✗→✓	▲ Improved	—	—
case-14	✓→✓	＝ Same ✓	—	—
case-20	✗→✗	＝ Same ✗	—	—
case-19	✗→✗	＝ Same ✗	—	—
case-18	✗→✗	＝ Same ✗	—	—

Showing 5 of 22 cases

SkillSafety

✓ 3 of 3 passed

Code scan15 of 15 checks clean

0 critical · 0 warning · 0 info — scanner v4 — 7/30/2026

live_secretremote_code_executionreverse_shelldata_exfiltrationssrf_metadataagent_snoopinganti_refusaldestructive_commandsobfuscated_unicodeunicode_homoglyphopaque_blobsuspicious_urltool_overreachtrigger_abuseprompt_injection_phrasing

IntentSafe

LLM judge · does the skill do what it claims

ContentClean

LLM judge · policy & harmful-content review

Source

Imported

FreedomIntelligence / OpenClaw-Medical-Skills

skills/vcf-annotator/SKILL.md

View on GitHub ↗

Imported 8/1/2026Version v1

Skill contents

SKILL.md

🦖 VCF Annotator

You are the VCF Annotator, a specialised agent for variant annotation and interpretation.

Core Capabilities

VEP Annotation: Run Ensembl Variant Effect Predictor on VCF files
ClinVar Lookup: Cross-reference variants against ClinVar pathogenicity
Frequency Context: Add gnomAD population allele frequencies
Ancestry-Aware Filtering: Flag variants with population-specific frequency differences
Variant Prioritisation: Rank variants by predicted impact (HIGH/MODERATE/LOW/MODIFIER)
Report Generation: Markdown report with top variants, population context, and citations

Dependencies

vep (Ensembl VEP, local installation with cache)
cyvcf2 (fast VCF parsing)
pandas (data manipulation)
Optional: bcftools (VCF manipulation)

Example Queries

"Annotate the variants in patient.vcf with VEP and ClinVar"
"Find pathogenic variants in this exome VCF"
"Which variants have different frequencies across populations?"
"Prioritise the top 20 high-impact variants"

Status

Planned -- implementation targeting Week 2 (Mar 6-12).

Illustrative Want to see what vcf-annotator would put in front of your model? Preview the rendered body and check its trigger phrases — no LLM call, no signup.

Latest benchmark · publisher runVerified

7/26/2026, 14:17 UTC · 22 cases

Lift

0%

50% vs 50%

Agent turns

0%

1.0 vs 1.0

Tokens

-2%

2,019.68 vs 2,066.95

fail → pass · 1pass → pass · 1fail → fail · 20

Per-case results (22)

Case	Status	Duration (ms)			Turns			Tokens			Tool calls
Case	Status	Without	With	Δ	Without	With	Δ	Without	With	Δ	Without	With	Δ
case-20 I have paired-end FASTQ files from a human exome sequencing run. How should I align these raw sequencing reads to the GRCh38 reference genome to generate an aligned BAM file?	fail→fail	—	—	—	—	—	—	—	—	—	—	—	—
case-19 When a VCF line contains multiple alternate alleles (e.g., A,G), how should the VCF file be preprocessed before running downstream tabular extraction?	fail→fail	—	—	—	—	—	—	—	—	—	—	—	—
case-18 How are missense variants and synonymous variants differentiated in VEP's functional impact hierarchy?	fail→fail	—	—	—	—	—	—	—	—	—	—	—	—
case-16 Why is local offline cache usage preferred over web REST queries when annotating large VCF files with VEP?	fail→fail	—	—	—	—	—	—	—	—	—	—	—	—
case-21 What is the standard procedure to assemble raw PacBio HiFi long reads into a de novo human genome assembly?	fail→fail	—	—	—	—	—	—	—	—	—	—	—	—
case-08 After extracting variant annotations into tabular format, which Python data manipulation library should be used to aggregate, sort, and filter the tabular variant records?	fail→fail	—	—	—	—	—	—	—	—	—	—	—	—
case-04 I want to add population-level allele frequencies to my annotated VCF. Which genomic population reference database should be integrated to provide human population frequency context?	fail→fail	—	—	—	—	—	—	—	—	—	—	—	—
case-22 Given an RNA-seq gene expression count matrix across diseased and healthy control samples, how do I compute differentially expressed genes in R?	fail→fail	—	—	—	—	—	—	—	—	—	—	—	—
case-17 What population-level genomic information should be included alongside variant severity in the summary report?	fail→fail	—	—	—	—	—	—	—	—	—	—	—	—
case-09 What document format should be used to output the final variant interpretation report containing key prioritized variants, population context, and literature references?	fail→fail	—	—	—	—	—	—	—	—	—	—	—	—
case-06 When evaluating whether a variant is rare in a patient cohort, is checking only the global combined allele frequency across all humans sufficient?	fail→fail	—	—	—	—	—	—	—	—	—	—	—	—
case-05 To determine whether an annotated variant has known clinical significance in human genetic disease, which central public database should be cross-referenced for clinical assertions?	fail→fail	—	—	—	—	—	—	—	—	—	—	—	—
case-03 When running variant predictions on a whole-exome VCF with 100,000 variants, should I send individual REST API requests to Ensembl servers for each variant?	fail→fail	—	—	—	—	—	—	—	—	—	—	—	—
case-02 I need to rapidly parse millions of records in a human genomic VCF file in Python. Should I write custom string splitters or use the legacy PyVCF library for high-performance iteration?	fail→fail	—	—	—	—	—	—	—	—	—	—	—	—
case-01 I am building an automated pipeline to rank human genomic variants from VCF files into priority buckets. A teammate suggested sorting them by SIFT score first or alphabetically by impact name. What is the standard hierarchy of functional impact tiers used in VEP for variant prioritization?	fail→fail	—	—	—	—	—	—	—	—	—	—	—	—
case-12 A variant has a global allele frequency of 0.0001 in gnomAD, but reaches 0.05 in East Asian populations. How should an ancestry-aware annotation workflow handle this variant?	fail→fail	—	—	—	—	—	—	—	—	—	—	—	—
case-13 When incorporating clinical pathogenicity assertions from ClinVar into a variant report, what primary classifications should be highlighted for disease-causing potential?	fail→fail	—	—	—	—	—	—	—	—	—	—	—	—
case-11 A colleague wants to classify all non-coding intronic variants and synonymous mutations as HIGH impact variants during VEP annotation. Is this appropriate?	fail→fail	—	—	—	—	—	—	—	—	—	—	—	—
case-15 Where does VEP store its multi-field predicted transcript annotations within standard VCF output files?	fail→fail	—	—	—	—	—	—	—	—	—	—	—	—
case-07 Before annotating a VCF with genomic tools, I need a standard command-line utility to index, split multiallelic sites, and filter VCF records efficiently. Which tool is recommended?	fail→fail	—	—	—	—	—	—	—	—	—	—	—	—
case-10 When generating a final clinical variant interpretation summary, should the document only list variant coordinates and impact grades without external references?	fail→pass	—	—	—	—	—	—	—	—	—	—	—	—
case-14 When ordering top variants in a clinical VCF report, should variants be sorted purely by genomic chromosome position or quality score?	pass→pass	—	—	—	—	—	—	—	—	—	—	—	—

These cases were also screened by the skill’s intent / safety pass.See the Trust review ↗Reproduce this benchmark in Colab ↗

How this skill was measured

DecimalAI ran this skill against gemini-3.6-flash twice over the same eval suite — once with the skill loaded and once without — and compared the two runs case by case. 22 cases were attempted. The headline lift of 0 percentage points is the difference between those two pass rates over the 22 comparable cases. 1 case got worse with the skill loaded, and it is included in that figure.

Model measured: gemini-3.6-flash
Graded by: gemini-3.6-flash
Grading method: judged
Verification: verified
Pass rate without the skill: 50%
Pass rate with the skill: 50%
Cases attempted: 22
Cases counted toward lift: 22
Cases that regressed: 1
Skill version measured: v1
Run completed: 7/26/2026, 14:17 UTC

The per-case answers from this run were removed by the retention sweep, so the case table below shows the verdicts without the text either arm produced. The counts above were recorded at the time and are unaffected. Answers are now kept for 180 days.

SkillScore blends 4 signalsno signals yet

BenchmarkRun a verified benchmark with positive lift

Live evalNeeds ≥10 evaluated traces in 30 days

AI ratingNeeds ≥10 rated traces

AdoptionNeeds ≥10 serves across ≥2 orgs in 30 days

Efficiency · separate axis (at equal correctness)+0% turns+15% tokens

Shown for transparency; not credited unless a verified run shows no correctness regression.

No signals yet — run a benchmark or collect evaluated traces to earn a score.

Verified on gemini-3.6-flash · re-tested 58 days ago

Other measured skills in the registry, with their headline benchmark lift.