methodsmith plugin

• @agent-methodsmith:method-administrative-data-reviewerinherit

  Conditional code-review persona, selected when the diff touches claims, billing, or payer data (Medicare, Medicaid, MarketScan, Optum). Reviews enrollment, look-back windows, coverage censoring, and coding drift.

• @agent-methodsmith:method-adversarial-document-reviewerinherit

  Conditional document-review persona, selected for documents with >5 requirements or units, major architecture decisions, high-stakes domains, or new abstractions. Challenges premises and unstated assumptions.

• @agent-methodsmith:method-adversarial-reviewerinherit

  Conditional code-review persona, selected for large diffs (>=50 changed lines) or high-risk areas (auth, payments, data mutations, external APIs). Constructs failure scenarios to break the implementation.

• @agent-methodsmith:method-architecture-strategistinherit

  Analyzes code changes from an architectural perspective for pattern compliance and design integrity. Use when reviewing PRs, adding services, or evaluating structural refactors.

• @agent-methodsmith:method-best-practices-researcherinherit

  Researches and synthesizes external best practices, documentation, and examples for any technology or framework. Use when you need industry standards, community conventions, or implementation guidance.

• @agent-methodsmith:method-bioinfo-pipeline-reviewerinherit

  Conditional code-review persona, selected when the diff touches Snakemake, Nextflow, CWL, Bioconductor, or sample-sheet pipelines. Reviews environment pinning, checksums, sample-sheet checks, and version traceability.

• @agent-methodsmith:method-calibration-reviewerinherit

  Conditional code-review persona for prediction-model evaluation code that outputs predicted risks (binary or survival). Use when a diff computes or reports discrimination, calibration, Brier score, or decision curves.

• @agent-methodsmith:method-causal-inference-reviewerinherit

  Conditional code-review persona, selected when the diff estimates causal effects from non-randomized data (propensity scores, IPTW, g-methods, target trial emulation, IV, DiD). Reviews DAG, confounders, positivity, and time zero.

• @agent-methodsmith:method-code-simplicity-reviewerinherit

  Reviews finished code for YAGNI violations and needless complexity, proposing the simplest working version. Use after implementation, before review.

• @agent-methodsmith:method-coherence-reviewerhaiku

  Reviews planning documents for internal consistency -- contradictions between sections, terminology drift, structural issues, and ambiguity where readers would diverge. Spawned by the document-review skill.

• @agent-methodsmith:method-concept-drift-reviewerinherit

  Conditional code-review persona, selected when the diff touches concept sets or ICD/CPT/LOINC/SNOMED code lists used across data refreshes or years. Detects vocabulary drift (ICD-9 to ICD-10, deprecated concepts).

• @agent-methodsmith:method-correctness-reviewerinherit

  Always-on code-review persona. Reviews code for logic errors, edge cases, state management bugs, error propagation failures, and intent-vs-implementation mismatches.

• @agent-methodsmith:method-data-leakage-reviewerinherit

  Conditional code-review persona, selected when the diff touches ML training or evaluation code (sklearn, xgboost, pytorch, tidymodels, survival ML). Detects target leakage, train-test contamination, and look-ahead bias.

• @agent-methodsmith:method-data-mapping-reviewerinherit

  Conditional code-review persona, selected when the diff touches a codebook, data dictionary, or SAP variable list. Flags missing variables, name drift, type mismatches, and unmapped columns before modeling.

• @agent-methodsmith:method-design-lens-reviewersonnet

  Reviews planning documents for missing design decisions -- information architecture, interaction states, user flows, and AI slop risk. Uses dimensional rating to identify gaps. Spawned by the document-review skill.

• @agent-methodsmith:method-fairness-reviewerinherit

  Conditional code-review persona, selected when the diff touches clinical prediction-model code. Reviews subgroup performance reporting (sex, race/ethnicity, age, site, payer) and flags gaps, omitted attributes, and missing mitigation.

• @agent-methodsmith:method-feasibility-reviewerinherit

  Evaluates whether proposed technical approaches in planning documents will survive contact with reality -- architecture conflicts, dependency gaps, migration risks, and implementability. Spawned by the document-review skill.

• @agent-methodsmith:method-framework-docs-researcherinherit

  Gathers comprehensive documentation and best practices for frameworks, libraries, or dependencies. Use when you need official docs, version-specific constraints, or implementation patterns.

• @agent-methodsmith:method-git-history-analyzerinherit

  Performs archaeological analysis of git history to trace code evolution, identify contributors, and understand why code patterns exist. Use when you need historical context for code changes.

• @agent-methodsmith:method-issue-intelligence-analystinherit

  Fetches and analyzes GitHub issues to surface recurring themes, pain patterns, and severity trends. Use when understanding a project's issue landscape, analyzing bug patterns for ideation, or summarizing what users are reporting.

• @agent-methodsmith:method-kieran-python-reviewerinherit

  Conditional code-review persona, selected when the diff touches Python code. Reviews changes with Kieran's strict bar for Pythonic clarity, type hints, and maintainability.

• @agent-methodsmith:method-learnings-researcherinherit

  Searches docs/solutions/ for past learnings that apply to the current task, matched by frontmatter metadata. Use before implementing, deciding, or starting work in a documented area.

• @agent-methodsmith:method-maintainability-reviewerinherit

  Always-on code-review persona. Reviews code for premature abstraction, unnecessary indirection, dead code, coupling between unrelated modules, and naming that obscures intent.

• @agent-methodsmith:method-methods-reviewerinherit

  Conditional review persona for inferential code and model-strategy artifacts. Reviews estimand alignment, model-family selection, mixed/random effects, clustering, marginalization, assumptions, and confounder handling.

• @agent-methodsmith:method-multiplicity-reviewerinherit

  Always-on review persona. Reviews multiple comparisons handling and bias indicators including endpoint multiplicity, correction methods, selective reporting, unplanned subgroup analyses, and p-hacking signals.

• @agent-methodsmith:method-omics-batch-reviewerinherit

  Conditional code-review persona, selected when the diff touches omics count or beta matrices and downstream analysis. Detects batch-condition confounding, unchecked ComBat/RUV/SVA use, and missing batch screening.

• @agent-methodsmith:method-omop-mapping-reviewerinherit

  Conditional code-review persona, selected when the diff touches OMOP CDM tables, concept_id columns, or concept sets. Reviews vocabulary pinning, validity dates, domains, descendants, and standard vs source concepts.

• @agent-methodsmith:method-pattern-recognition-specialistinherit

  Analyzes code for design patterns, anti-patterns, naming conventions, and duplication. Use when checking codebase consistency or verifying new code follows established patterns.

• @agent-methodsmith:method-performance-oracleinherit

  Analyzes code for performance bottlenecks, algorithmic complexity, database queries, memory usage, and scalability. Use after implementing features or when performance concerns arise.

• @agent-methodsmith:method-performance-reviewerinherit

  Conditional code-review persona, selected when the diff touches database queries, loop-heavy data transforms, caching layers, or I/O-intensive paths. Reviews code for runtime performance and scalability issues.

• @agent-methodsmith:method-phi-leak-reviewerinherit

  Conditional code-review persona, selected when the diff touches data files, codebooks, notebooks, manuscripts, or rendered artifacts. Detects HIPAA Safe Harbor identifiers leaking into the repository.

• @agent-methodsmith:method-pr-comment-resolverinherit

  Evaluates and resolves one or more related PR review threads -- assesses validity, implements fixes, and returns structured summaries with reply text. Spawned by the resolve-pr-feedback skill.

• @agent-methodsmith:method-previous-comments-reviewerinherit

  Conditional code-review persona, selected when reviewing a PR that has existing review comments or review threads. Checks whether prior feedback has been addressed in the current diff.

• @agent-methodsmith:method-product-lens-reviewerinherit

  Reviews planning documents as a senior product leader: challenges premises, weighs strategic consequences and opportunity cost, and surfaces goal-work misalignment. Spawned by the document-review skill.

• @agent-methodsmith:method-project-standards-reviewerinherit

  Always-on code-review persona. Audits changes against the project's designated, path-scoped criteria files and cites each violated rule.

• @agent-methodsmith:method-python-ds-reviewerinherit

  Conditional code-review persona for Python data science code. Reviews pandas, sklearn pipelines, leakage, statsmodels usage, and mixed-model/GEE implementation boundaries.

• @agent-methodsmith:method-quarto-render-reviewerinherit

  Conditional code-review persona, selected when the diff touches .qmd, _quarto.yml, or rendered output. Reviews Quarto/R Markdown for cache traps, render-only failures, parameter gaps, and committed-output hygiene.

• @agent-methodsmith:method-r-code-reviewerinherit

  Conditional code-review persona, selected when the diff touches R code. Reviews R code for data science quality -- tidyverse/base consistency, dplyr patterns, pipe usage, ggplot2, data.table, purrr, and R-specific anti-patterns.

• @agent-methodsmith:method-r-pipeline-reviewerinherit

  Conditional code-review persona, selected when the diff touches R analysis pipelines. Reviews statistical correctness: dplyr logic, survival and mixed-model patterns, convergence, ggplot2 accessibility.

• @agent-methodsmith:method-reliability-reviewerinherit

  Conditional code-review persona, selected when the diff touches error handling, retries, circuit breakers, timeouts, health checks, background jobs, or async handlers. Reviews code for production reliability and failure modes.

+15

• method-adversarial-review

  Challenges high-stakes code diffs or plans using adversarial analysis and cross-model peer reviews to identify failure modes and assumptions.

• method-anti-slop

  Detects and removes generic AI-writing tells from scientific manuscripts while preserving necessary passive voice in Methods sections.

• method-brainstorm

  Explores research ideas, compares study designs, and defines requirements to create a protocol or analysis brief for future planning.

• method-checklist

  Selects the appropriate EQUATOR reporting checklists for research studies based on design, AI involvement, and population.

• method-citations

  Builds evidence-traceable citations from PubMed and Paperclip into a DOCX with editable Zotero fields for manuscript auditing and repair.

• method-clean-branches

  Deletes local branches and associated worktrees whose remote tracking branches no longer exist after user confirmation.

• method-clif

  Enforces Common Longitudinal ICU data Format rules including mCIDE versions, Parquet usage, UTC datetimes, and PHI protection for ICU data analysis.

• method-clif-audit

  Audits CLIF project repositories for compliance with consortium conventions, reproducibility, and data safety without modifying files.

• method-clif-meta-analysis

  Plans and executes formal R or Python meta-analyses for pooling CLIF site effects or conducting systematic literature reviews.

• method-clif-template

  Initializes or audits a CLIF analysis repository to ensure structural consistency, PHI safeguards, and compliance with official project templates.

• method-clifpy

  Writes and reviews Python code using the clifpy library for CLIF data processing, ensuring correct API usage, versioning, and timezone handling.

• method-cohort

  Defines reproducible research cohorts by generating concept sets, SQL queries, and CONSORT-style waterfall counts for EHR, OMOP, or claims data.

• method-commit

  Creates focused conventional git commits while automatically excluding raw data, PHI, and large files from version control.

• method-compound

  Documents validated analytical approaches, debugging decisions, or data science patterns into structured solution files.

• method-compound-refresh

  Audits documentation and solution learnings for drift to update, merge, replace, or mark them as stale.

• method-data-qa

  Profiles data columns and executes quality gates to validate extracts against SAP expectations before modeling or unblinding.

• method-debug

  Reproduces, traces, and fixes failing code, pipelines, or tests by identifying the root cause through a causal chain analysis.

• method-doc-review

  Reviews plans, protocols, and manuscripts for clarity and completeness using specialized reviewer personas.

• method-evidence-map

  Builds a traceable evidence map for biomedical research questions using PubMed metadata and optional Paperclip full-text deepening.

• method-extract

  Extracts structured methods, cohorts, endpoints, and statistical models from PubMed papers to provide grounding for Statistical Analysis Plans.

• method-fabric

  Routes Microsoft Fabric tasks to specific execution engines like PySpark, Data Factory, or KQL while applying research guardrails.

• method-figure

  Creates and validates publication-quality clinical figures using ggplot2 or matplotlib with standardized styling and manifest tracking.

• method-genome-build

  Pins reference genome builds, annotation versions, and contig naming to ensure consistency across bioinformatics datasets.

• method-grill

  Interviews researchers one question at a time to stress-test study goals, hypotheses, data, and analysis plans before submission.

• method-handoff

  Creates and resumes portable continuity records for data science tasks to preserve context between different agents or sessions.

• method-ideate

  Generates and ranks grounded research or analysis ideas based on repository, literature, or topic context.

• method-journal-select

  Matches manuscripts or abstracts to candidate journals by analyzing scope fit, indexing, and access policies to provide a ranked submission list.

• method-language-detect

  Detects R, Python, or mixed-language signals in repository files to provide language handoff data for downstream skills.

• method-lit-search

  Routes literature requests to specialized tools like PubMed, full-text evidence mapping, or authoritative web sources based on the research need.

• method-marimo

  Creates, edits, validates, and exports reactive Python notebooks using text-native .py files and UI widgets.

• method-meta-analysis

  Pools literature effect sizes using REML and Hartung-Knapp methods to provide anchored effect-size assumptions for power calculations.

• method-ml-track

  Adds reproducible machine learning experiment tracking using MLflow, W&B, DVC, or offline logs to capture data hashes, metrics, and validation metadata.

• method-model-card

  Generates clinical model documentation including intended use, performance metrics, calibration, fairness, and limitations based on evaluation data.

• method-model-choice

  Selects and justifies statistical models like regression, survival, or Bayesian methods based on research aims, data structure, and estimands.

• method-notebook-edit

  Edits existing Jupyter notebooks using cell anchors and backups while preserving metadata and JSON structure.

• method-notebook-standards

  Enforces structural and documentation standards for Marimo, Jupyter, R Markdown, and Quarto notebooks to ensure reproducibility and reviewability.

• method-numbers

  Locks manuscript numbers to an analysis registry to ensure every reported result traces back to its specific script commit, source hash, and SAP output.

• method-omics-qc

  Performs quality control for sequencing and omics data including FASTQ, BAM, VCF, and methylation arrays to detect sample swaps and batch confounding.

• method-optimize

  Runs bounded, metric-scored experiment loops to improve models, prompts, or pipelines using cross-validation and experiment tracking.

• method-phenotype

  Validates a computable phenotype against chart review: 2x2 table, PPV, NPV, sensitivity, specificity, Wilson intervals. Use when checking a phenotype or code list.

+30

• PreToolUsecommand

  Read|Grep|Bash|Edit|MultiEdit|NotebookRead|NotebookEdit|Moni