The Agent Skills Directory

[SAFE]: The skill implements standard bioinformatics workflows using widely-accepted scientific libraries and practices. All components are transparent and well-documented.
[EXTERNAL_DOWNLOADS]: The skill fetches gene-set data and performs analysis using official APIs from trusted scientific organizations.
Fetches data from Enrichr (Ma'ayan Lab), g:Profiler (University of Tartu), and MSigDB (Broad Institute).
Utilizes the gseapy and gprofiler-official Python packages, which are reputable tools in the bioinformatics community.
[INDIRECT_PROMPT_INJECTION]: The skill processes external data such as gene lists and expression matrices, which is its primary purpose.
Ingestion points: Files read via scripts/run_enrichment.py (--genes, --deseq2, --rnk, --background) and gene lists passed to gseapy functions.
Boundary markers: None explicitly present for instruction isolation, but data is treated as structured bioinformatics input.
Capability inventory: Network requests to enrichment APIs and local file system writes for results and plots.
Sanitization: scripts/run_enrichment.py performs gene symbol normalization, deduplication, and numeric coercion of scores.
[COMMAND_EXECUTION]: The provided Python script scripts/run_enrichment.py performs legitimate, documented data analysis tasks. It does not use unsafe execution functions like eval() or exec(), nor does it spawn unauthorized shell commands.

pathway-enrichment