Browse reusable Agent Skills, each with a clear purpose and practical guidance.
Load when computing ligand-receptor cell-cell communication on a preprocessed spatial AnnData with `obs[cell_type_key]` (default `leiden`) via LIANA (default), CellPhoneDB, FastCCC, or CellChat (R). Skip when running scRNA-only L-R inference (use `sc-cell-communication`) or when no cell-type labels exist (run `spatial-annotate` or `spatial-domains` first).
Load when ranking spatial cluster markers or comparing two spatial groups in spatial transcriptomics. Skip if the data is single-cell (use sc-de) or bulk (use bulkrna-de), or for spatially variable expression discovery (use spatial-genes).
Load when detecting tissue domains / niches on a preprocessed spatial AnnData via Leiden / Louvain (spatial-weighted) or graph-neural backends (SpaGCN / STAGATE / GraphST / BANKSY / CellCharter). Skip when ranking spatially variable genes (use spatial-genes) or for spot-level cell-type annotation (use spatial-annotate).
Load when running pathway / gene-set enrichment per cluster on a preprocessed spatial AnnData via Enrichr (over-representation), GSEA (preranked), or ssGSEA (per-cell scores). Skip when ranking spatially variable genes (use `spatial-genes`) or when comparing pathways across conditions (use `spatial-condition` for DE first, then this skill on the ranked output).
Fast, cloud-based patent searching across 100 million+ worldwide patents using Google BigQuery - keyword search, CPC classification, patent details retrieval
Load when querying STRING for the protein-protein interaction subgraph induced by a bulk RNA-seq DEG list and finding hub genes. Skip for pathway enrichment of the same list (use bulkrna-enrichment) or for de novo co-expression network discovery (use bulkrna-coexpression).
Load when placing bulk RNA-seq samples on a single-cell reference's pseudotime axis (NNLS deconvolution + nearest-neighbour mapping). Skip for plain cell-type proportions (use bulkrna-deconvolution alone) or for native single-cell trajectory inference (use sc-pseudotime).
Production-grade Bayesian causal inference with PyMC, CausalPy, and DoWhy. Enforces DAG-first thinking, mandatory user checkpoints for assumptions, design-specific refutation, and defensible reporting with causal language guardrails. Trigger on: causal inference, causal effect estimation, treatment effects, counterfactuals, difference-in-differences (DiD), synthetic control, regression discontinuity (RDD), interrupted time series (ITS), instrumental variables (IV), propensity scores, DAGs, causal graphs, confounders, backdoor criterion, do-calculus, interventional distributions, pm.do(), pm.observe(), CausalPy, DoWhy, mediation analysis, refutation, sensitivity analysis, parallel trends, placebo tests, or any question of the form "does X cause Y" or "what is the effect of X on Y."
Use when the user wants to list, search, or inspect available SDG flows and data generation pipelines. Applies to browsing flow catalogs, finding flows by use case, or understanding what a specific flow does.
Load when running over-representation analysis (ORA) on a metabolite list via Fisher's exact test against a built-in 9-pathway DEMO dictionary, BH-FDR adjusted. Skip when needing real KEGG / Reactome (this skill is demo-only) or `mummichog` / `fella` topology methods (CLI accepts them but only ORA runs).