reproducibility plugins

by Manuel Corpas

Bioinformatics-native AI agent skill library: pharmacogenomics, clinical variant interpretation, GWAS and polygenic scores, ancestry, RNA-seq and single-cell, metagenomics, proteomics and nf-core pipeline wrappers. 97 Agent Skills, 50 with a deterministic CLI entry point and demo data, every runnable skill writing a SHA-256 reproducibility bundle. Local-first.

by aiopshwang

Evidence-first AI agent skills for reliable data analysis, machine learning, and reproducible model validation.

by Matt Corbett

Bioinformatics & genomics-pipeline team — agents (bioinformatics-workflow-architect, genomics-pipeline-engineer) for the layer answering 'which workflow engine, reference, and compute strategy, and how do we build a reproducible, validated genomics pipeline?': engine choice (Nextflow/nf-core, Snakemake, WDL+Cromwell/miniwdl, CWL), reference build (GRCh38 vs T2T-CHM13) + core steps (QC, trimming, alignment BWA-MEM2/minimap2/STAR/Salmon, dedup, variant calling GATK/DeepVariant, joint genotyping), RNA-seq (DESeq2/edgeR) and single-cell (Scanpy/Seurat), reproducibility (Docker/Apptainer, Conda/Bioconda, pinned versions, FAIR, provenance), HPC-vs-cloud scaling and cost, and validation vs GIAB/hap.py truth sets. skills, a knowledge bank (pipeline decision tree + 2026 workflow-patterns reference), and templates. Distinct from ml-engineering (generic MLOps/model training), clinical-trials (trial operations/regulatory), and data-platform (warehouse/BI). Requires ravenclaude-core@>=0.7.0.

by Matt Corbett

Exploratory data science & reproducible research team — agents (exploratory-data-scientist, feature-and-modeling-engineer, research-reproducibility-engineer) for the analysis layer between raw data and a defensible result: data profiling/cleaning, EDA, hypothesis generation, communicating findings with uncertainty; feature engineering, classical modeling (regression, trees, boosting), model selection, evaluation (cross-validation, metrics) and leakage avoidance; and the reproducibility spine — notebook hygiene, pinned environments, experiment tracking, data/version control, seeds, and reproducible pipelines. best-practices, decision-tree knowledge bank, skills, commands, templates, a hook, scenarios bank, and stdlib-only ds_calc.py (classification/regression metrics, split-check). Seams: is-the-result-statistically-real -> applied-statistics; productionize/serve/monitor -> ml-engineering; data pipelines/warehouse -> data-platform. Requires ravenclaude-core@>=0.7.0.

by skymanbp

Scientific paper writing, source-traced review, and de-AI rewriting for top-tier journals (ApJ/MNRAS/PRD/JCAP-class) and NSF/NIH proposals. Ships 12 skills and 40 tools under one typed standard: scientific-integrity blockers that cannot be waived as style preferences, a zero-target L0 lexicon, ranked advisories, explicit measurement states so missing calibration never reads as clean, claim-first rewriting gated on hard scientific-fidelity eligibility, a mechanical length budget with a machine-built removal map and a required-shrink gate, corpus-calibrated salience-hierarchy, domain-register (with an exhaustive zero-hit vocabulary audit), collocation, and discourse-texture (given/new cohesion, epistemic hedging) axes, each judged against the field's own per-section distribution at the unit it was calibrated at, a residue axis that gates the trace an edit leaves, A-R review with narrative-spine and adversarial-escalation dimensions, registry-verified bibliographies (CrossRef, DataCite, arXiv), pixel-measured figure review, and isolated multi-reviewer orchestration. Learned scores are field-similarity triage, never an authorship verdict. See CHANGELOG.md for per-version history.

by David Kim

Verification loop for analytics and research deliverables: claim ledger, mechanical check script, independent verifier on a fresh context, re-check of every fix, and a receipt bound to the exact file sent.

by Open Science Pillars Community

Foundation: earth science data formats, statistics, uncertainty quantification, cartography, quality control, reproducibility, analysis review.

by Hodios contributors

Hodios ai-ml entries: prompts, personas and workflows.

by Hodios contributors

Hodios devops entries: prompts, personas and workflows.

by Hodios contributors

Hodios statistics entries: prompts, personas and workflows.

by Basile Chrétien

Make every number in a scientific manuscript traceable to its source. Skills for setting up a paper project, drafting, literature verification, figure review, journal compliance, review panels, the Word round trip and the response to reviewers; hooks that refuse edits to machine-written files and hold a submission to submission standards.

by pliperkiller

Rules and workflows that teach a coding agent to produce scientific research output: figures, notebooks, papers, slides and analysis scripts.

by Anna Rosen

Research-coding workflow for computational science, built with a JAX/Python astrophysics group and domain-agnostic in its contracts. The researcher owns the science: scientific assumptions are proposed and approved before they become load-bearing, experiments carry a derived prediction, and validation is staged (physics and analytic checks in development, published-result or reference-code comparison before release). 43 skills: model development and ideation, literature, scoping, building correctly, equation-critical sources, verification and numerical validation, Bayesian inference, code and figure review, performance, records and provenance, MyST documentation, reproducibility and release. Plus an equation-verifier agent, six slash commands, twelve self-limiting hooks (Stop gates warn by default; RWF_STRICTNESS=standard blocks), and per-domain lenses for reference-parity work. The same skill files load in Codex via a symlink.