The library

screenspot-param-sweep · Best for adversarial evaluation across agent personas.

classify-improve · improving classifiers via data-driven signal mining from real failures

scout-and-combine · fan-out parallel research with synthesized results

rebench-binary-survey · Comprehensive multi-phase scan across heterogeneous source corpus with parallel validation.

model-convert-benchmark · Multi-phase model-convert-benchmark orchestration with structured phase control and receipts.

references-audit-classify · Comprehensive multi-phase scan across heterogeneous source corpus with parallel validation.

cf-parallel-harvester-rawload · Append-loading UK Contracts Finder historical data (2016-2026) into BigQuery in resumable 2-month shards.

news-graph-extract · Building knowledge graphs from news corpora where entities and relationships must be structured for Neo4j ingestion.

SKILL · Automated data quality gates in ML pipelines with both fast rule checks and semantic LLM scoring

trader-risk · Risk management for traders when you need VaR, position sizing, and circuit breaker alerts in one command.

sqlite-incremental-reconcile · When sqlite-incremental-reconcile is needed before shipping or in production.

server-pdf · When server-pdf is needed before shipping or in production.

evaluate-model · Standardized model evaluation reporting post-training.

generate-report · Creating shareable experiment summaries for stakeholder review.

run-pipeline · Reproducible end-to-end execution after configuration or code changes.

firecrawl-mcp · When large-scale web scraping or extraction is needed beyond simple HTTP fetch.

postgres-mcp · When agents need direct database access for queries or mutations.

cheap-summarizer · When this specialized capability is required.

github-repo · When this specialized capability is required.

site-status · When this specialized capability is required.

trending · When this specialized capability is required.

usdc-summarize · When this specialized capability is required.

whois-lookup · Verifying domain ownership history, checking expiration dates, or extracting registrar and nameserver info for domain takeover assessment.

modal · Running heavy compute (ML training, data processing), GPU-dependent code, or untrusted code in safe isolation without local GPU procurement or Docker management.

prediction-markets · Getting real-time event probabilities and comparing prices across venues

rpc · Low-level blockchain reads where native library integration is overkill

surf · Quick wallet/contract state snapshots without running full indexing

query-onchain-data · Querying decoded blocks, transactions, and events on Base via standardized SQL without running a local full node.

ultraresearch · When you need maximum-saturation research across every available source.

airflow-dag-patterns · Teams automating repetitive tasks while maintaining control over execution and output quality.

after_skill · Automating post-work tasks like syncing, cleanup, or notifications

backtesting-frameworks · Validating trading strategies when look-ahead bias, survivorship bias, and transaction costs must be accounted for.

benchmark · Catching performance regressions in real browser metrics before and after code changes or before shipping.

agent-browser · Automating form submission, data extraction, and web app testing when refs are more resilient than XPath.

brightdata · Sites with bot detection or CAPTCHA that block naive fetch

aeon · Temporal pattern detection faster/more accurate than standard ML

anndata · Single-cell genomics where scverse integration is mandatory

astropy · Astronomy research when specialized units + coordinates matter

benchling-integration · Biotech workflows needing certified ELN audit trails

bgpt-paper-search · Biotech literature review when Google Scholar insufficient

bids · When organizing multi-modal neuroimaging/biomedical datasets for repositories (OpenNeuro, DANDI) or validating BIDS compliance.

biopython · When writing batch molecular biology pipelines that manipulate sequences, parse multiple file formats, and access NCBI programmatically.

bioservices · When integrating data from multiple bioinformatics databases (UniProt, KEGG, ChEMBL) in a single workflow with consistent ID mapping.

bulk-rnaseq · When analyzing bulk RNA-seq from raw reads through differential expression and pathway enrichment with reproducible, quality-gated pipelines.

cellxgene-census · When querying organism-wide single-cell expression data across tissues without downloading multi-terabyte datasets.

cirq · When designing and simulating quantum algorithms before submission to real quantum hardware.

dask · Scaling existing pandas code to multi-GB datasets without Spark rewrite.

database-lookup · One-query lookups across chemistry, biology, regulatory databases.

datamol · Drug discovery pipelines avoiding low-level RDKit chemistry code.

deepchem · End-to-end neural property prediction when graph models beat hand features.