ml-research plugin
Machine learning experiments, training, evaluation and reproducible papers.
How to install the ml-research plugin
Run both commands inside Claude Code. The first adds the marketplace once; the second installs the plugin.
/plugin marketplace add KalarisLabs/research-agent-skills
/plugin install ml-research@research-agent-skillsWhat's inside ml-research
12 parts read from the plugin's own files.
Skills 12
Claude loads it when the task matches- academic-plotting
Generates publication-quality figures for ML papers from research context. Given a paper section or description, extracts system components and relationships to generate architecture diagrams via Gemini. Given experiment results or data, auto-selects chart type and generates data-driven figures via matplotlib/seaborn. Use when creating any figure for a conference paper.
- evaluating-llms-harness
Evaluates LLMs across 60+ academic benchmarks (MMLU, HumanEval, GSM8K, TruthfulQA, HellaSwag). Use when benchmarking model quality, comparing models, reporting academic results, or tracking training progress. Industry standard used by EleutherAI, HuggingFace, and major labs. Supports HuggingFace, vLLM, APIs.
- exploratory-data-analysis
Perform bounded, local exploratory analysis of explicitly supported scientific files. Use for redacted CSV/TSV/JSON profiles; optional NumPy, HDF5, FASTA/FASTQ, and basic image metadata inspection; missingness/leakage audits; outlier and transformation sensitivity; and rigorous EDA report scaffolds. Other domain formats are reference-only and unknown formats fail closed.
- huggingface-accelerate
Wraps existing PyTorch training scripts with HuggingFace Accelerate (Accelerator class, accelerate config, accelerate launch) so the same code runs on CPU, single GPU, multi-GPU, multi-node, TPU, or Apple MPS. Covers device placement, FP16/BF16/FP8 mixed precision, gradient accumulation, distributed checkpointing, and switching between DDP, DeepSpeed ZeRO, FSDP, and Megatron backends. Use when converting a single-GPU script to multi-GPU, enabling mixed precision, configuring DeepSpeed ZeRO or FSDP, or setting up gradient accumulation. Use when one script must run on different hardware. Not for callback-heavy training loops (use PyTorch Lightning) or multi-node orchestration with hyperparameter tuning (use Ray Train).
- ml-paper-writing
No description in the file.
- ml-training-recipes
No description in the file.
- pytorch-lightning
No description in the file.
- reproducibility-statement
No description in the file.
- scikit-learn
No description in the file.
- shap
No description in the file.
- statistical-analysis
No description in the file.
- transformers
No description in the file.
Using ml-research in Claude Code
Skills need no command. Claude reads them when your request matches what they describe.
Update, turn off or remove
- Update the plugin (shell)
- claude plugin update ml-research@research-agent-skills
- Refresh the marketplace
- /plugin marketplace update research-agent-skills
- Turn it off for now
- /plugin disable ml-research@research-agent-skills
- Remove it
- /plugin uninstall ml-research@research-agent-skills
Skills in ml-research
- exploratory-data-analysisKalarisLabs/research-agent-skills
Perform bounded, local exploratory analysis of explicitly supported scientific files. Use for redacted CSV/TSV/JSON profiles; optional NumPy, HDF5, FASTA/FASTQ, and basic image metadata inspection; missingness/leakage audits; outlier and transformation sensitivity; and rigorous EDA report scaffolds. Other domain formats are reference-only and unknown formats fail closed.
21 files126stars - academic-plottingKalarisLabs/research-agent-skills
Generates publication-quality figures for ML papers from research context. Given a paper section or description, extracts system components and relationships to generate architecture diagrams via Gemini. Given experiment results or data, auto-selects chart type and generates data-driven figures via matplotlib/seaborn. Use when creating any figure for a conference paper.
4 files126stars - evaluating-llms-harnessKalarisLabs/research-agent-skills
Evaluates LLMs across 60+ academic benchmarks (MMLU, HumanEval, GSM8K, TruthfulQA, HellaSwag). Use when benchmarking model quality, comparing models, reporting academic results, or tracking training progress. Industry standard used by EleutherAI, HuggingFace, and major labs. Supports HuggingFace, vLLM, APIs.
7 files126stars - huggingface-accelerateKalarisLabs/research-agent-skills
Wraps existing PyTorch training scripts with HuggingFace Accelerate (Accelerator class, accelerate config, accelerate launch) so the same code runs on CPU, single GPU, multi-GPU, multi-node, TPU, or Apple MPS. Covers device placement, FP16/BF16/FP8 mixed precision, gradient accumulation, distributed checkpointing, and switching between DDP, DeepSpeed ZeRO, FSDP, and Megatron backends. Use when converting a single-GPU script to multi-GPU, enabling mixed precision, configuring DeepSpeed ZeRO or FSDP, or setting up gradient accumulation. Use when one script must run on different hardware. Not for callback-heavy training loops (use PyTorch Lightning) or multi-node orchestration with hyperparameter tuning (use Ray Train).
4 files126stars
ml-research plugin questions
How do I install the ml-research plugin?
In Claude Code run /plugin marketplace add KalarisLabs/research-agent-skills, then /plugin install ml-research@research-agent-skills.
What does the ml-research plugin include?
ml-research bundles 12 skills.
How do I update or remove the ml-research plugin?
Update it with claude plugin update ml-research@research-agent-skills in your shell, or from the Installed tab of /plugin. Remove it with /plugin uninstall ml-research@research-agent-skills.
What license is the ml-research plugin under?
MIT, as stated in the plugin's manifest.