labrat

Plugin labrat from marketplace claude-settings in fcakyon/claude-codex-settings. Commands read from the repo's manifest — today.

Skills (11)

• compare

  Same-epoch comparison and analysis of training runs across wandb, neptune, tensorboard, mlflow, or local result files. Aligns runs at the student's current step (never current-vs-final-of-baseline) and separates proxy metrics from downstream targets. Also writes run summaries and experiment reports,

• dataset-curation

  Use when the user wants to analyze dataset bias, create stratified samples, evaluate fairness, or plan dataset collection. Triggers on phrases like "dataset bias", "stratified sample", "class imbalance", "data distribution", "fairness analysis", or "ethical review".

• debug

  Evidence-before-action diagnosis of failing ML experiments. Probes the system before guessing causes, process list, dmesg, GPU stats, log scrollback, checkpoint state, then states a hypothesis as a hypothesis and runs a smoke before claiming a root cause. Use when the user asks why a run is failing,

• experiment-design

  Use when the user wants to design experiments, plan ablation studies, structure baselines, or create incremental evaluation strategies. Triggers on phrases like "design ablation", "plan experiment", "what experiments should I run", "baseline comparison", or "experiment matrix".

• launch

  Pre-flight checklist for long-running ML training jobs covering config diff, run naming, path verification, monitoring setup, and restart-cleanup. Use when the user asks to launch, kick off, start, restart, or kill a training run, or mentions launching a multi-hour or multi-day GPU job (python train

• literature-research

  Use when the user wants to find related work, survey a research area, run a literature gap analysis, or discover open-source implementations. Maps existing work on a topic and confirms with web search what is missing. Triggers on phrases like "find papers on", "related work", "literature review", "w

• paper-check

  Check a paper against its code, results and citations before submission (numbers, terms, code vs paper, formulas vs code, BibTeX and cited claims via DBLP and Crossref, eval integrity), and prepare its code for open-source release. Use when the user says "verify claims" / "verify paper against code"

• paper-write

  Write, structure, or revise academic paper sections, keep notation consistent, refine figures and tables, and set up or fix the LaTeX build (TeX distribution, biber vs bibtex, venue templates, compilation). Use when the user says "write the abstract", "structure the methods", "improve this section",

• rebuttal

  Prepare a paper for peer review. Reads the paper and experiment results, selects the strongest ablations, anticipates reviewer questions, finds weaknesses before submission, and drafts rebuttals to actual reviews. Use when the user asks about "reviewer questions", "anticipate reviewers", "rebuttal",

• reproduce

  End-to-end paper reproduction from arxiv URL through smoke runs to replication experiments. Handles missing or partial official code, missing training scripts, missing hyperparameters, and private datasets via similar-public-dataset substitution. Use when the user asks to reproduce, implement, repli

• setup

  30-second auto-detection tour, then optional configuration of notifications, CLI allowlist, research CLAUDE.md rules, and LaTeX environment. Use when the user runs /labrat:setup.

Similar plugins