fine-tuning-expert
Use when fine-tuning LLMs, training custom models, or adapting foundation models for specific tasks. Invoke for configuring LoRA/QLoRA adapters, preparing JSONL training datasets, setting hyperparameters for fine-tuning runs, adapter training, transfer learning, finetuning with Hugging Face PEFT, OpenAI fine-tuning, instruction tuning, RLHF, DPO, or quantizing and deploying fine-tuned models. Trigger terms include: LoRA, QLoRA, PEFT, finetuning, fine-tuning, adapter tuning, LLM training, model training, custom model.
94
92%
Does it follow best practices?
Impact
94%
1.01xAverage score across 6 eval scenarios
Advisory
Suggest reviewing before use
Quality
Discovery
100%Based on the skill's description, can an agent find and select it at the right time? Clear, specific descriptions lead to better discovery.
This is a strong skill description that clearly defines its domain (LLM fine-tuning), lists specific concrete actions, provides explicit 'Use when' guidance, and includes comprehensive trigger terms covering common user phrasings and technical variations. It uses proper third-person voice throughout and is well-structured for skill selection among many candidates.
| Dimension | Reasoning | Score |
|---|---|---|
Specificity | Lists multiple specific concrete actions: configuring LoRA/QLoRA adapters, preparing JSONL training datasets, setting hyperparameters, adapter training, transfer learning, quantizing and deploying fine-tuned models. | 3 / 3 |
Completeness | Clearly answers both 'what' (configuring adapters, preparing datasets, setting hyperparameters, quantizing/deploying) and 'when' with an explicit 'Use when' clause at the start and explicit trigger terms listed at the end. | 3 / 3 |
Trigger Term Quality | Excellent coverage of natural terms users would say, including common variations: LoRA, QLoRA, PEFT, finetuning, fine-tuning, adapter tuning, LLM training, model training, custom model, Hugging Face PEFT, OpenAI fine-tuning, instruction tuning, RLHF, DPO. | 3 / 3 |
Distinctiveness Conflict Risk | Highly distinctive niche focused specifically on LLM fine-tuning and adapter training. The specific trigger terms like LoRA, QLoRA, PEFT, DPO, RLHF are unlikely to conflict with other skills. | 3 / 3 |
Total | 12 / 12 Passed |
Implementation
85%Reviews the quality of instructions and guidance provided to agents. Good implementation is clear, handles edge cases, and produces reliable results.
This is a strong skill with excellent actionability — the complete, executable LoRA fine-tuning example with QLoRA and merge variants is highly practical. The workflow is well-structured with explicit validation checkpoints at each stage, and the progressive disclosure via the reference table is clean and well-organized. Minor conciseness improvements could be made by trimming the persona line and some redundant constraint items.
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | The skill is mostly efficient but includes some unnecessary framing (e.g., 'Senior ML engineer specializing in...') and the constraints section partially restates what's already implied by the workflow. The code example is well-commented but some comments explain things Claude would already know (e.g., 'effective batch size = 16'). Overall reasonably lean but could be tightened. | 2 / 3 |
Actionability | The skill provides a fully executable, copy-paste-ready LoRA fine-tuning example with real library imports, concrete hyperparameter values, and working code for QLoRA and adapter merging. The dataset format, training arguments, and deployment steps are all specific and executable. | 3 / 3 |
Workflow Clarity | The core workflow has 5 clearly sequenced steps, each with explicit validation checkpoints (dataset validation before training, validation loss monitoring during training, benchmark evaluation before deployment). There are clear feedback loops — fix errors before proceeding, plateau signals overfitting — and the constraints reinforce these checkpoints. | 3 / 3 |
Progressive Disclosure | The skill has a clean overview workflow, a reference table with 5 topic-specific files and clear 'Load When' triggers, and the main content stays at the right level of detail. References are one level deep and well-signaled with a structured table format. | 3 / 3 |
Total | 11 / 12 Passed |
Validation
100%Checks the skill against the spec for correct structure and formatting. All validation checks must pass before discovery and implementation can be scored.
Validation — 11 / 11 Passed
Validation for skill structure
No warnings or errors.
- Repository
- jeffallan/claude-skills
- Commit
3d95bb1
- Reviewed
Table of Contents
Is this your skill?
If you maintain this skill, you can claim it as your own. Once claimed, you can manage eval scenarios, bundle related skills, attach documentation or rules, and ensure cross-agent compatibility.