[New Skill]: Multi-Omics FAIR Normalizer

[rosspeili]

Skill Name

bioinformatics/omics_data_normalizer

What should this skill do?

Ideal for CERTH INAB's Biodata Analysis Group

Biological data is notoriously messy. When specialized AI agents are deployed to analyze genetic or transcriptomic CSVs, they often hallucinate relationships due to poor formatting. This skill accepts a raw genomic/multi-omics payload and automatically restructures it to adhere to FAIR principles (Findable, Accessible, Interoperable, Reusable). It outputs standardized metadata headers so the main LLM can reliably execute bioinformatics reasoning.

Contributors: This should strictly process the data structure, not run heavy genomic alignments. In instructions.md, clarify that this skill is for formatting and validation, not diagnosis. Ensure pandas is listed in the manifest.yaml requirements.

Ideal Inputs & Outputs

Input:
{
"raw_csv_path": "./local_data/patient_transcriptome_raw.csv",
"target_standard": "MIAME"
}

Output:
{
"status": "normalized",
"fair_compliant_json_path": "./local_data/patient_transcriptome_fair.json",
"missing_metadata_warnings": ["sample_collection_date"]
}

Targeted Models (if applicable)

Model Agnostic (All)