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.
90
88%
Does it follow best practices?
Impact
Pending
No eval scenarios have been run
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 concrete actions, provides explicit trigger guidance with 'Use when' and 'Invoke for' clauses, and includes a comprehensive list of natural trigger terms. It occupies a distinct niche that is unlikely to conflict with other skills. The description uses appropriate third-person voice throughout.
| 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, instruction tuning, RLHF, DPO. | 3 / 3 |
Completeness | Clearly answers both 'what' (configuring adapters, preparing datasets, setting hyperparameters, quantizing/deploying models) and 'when' with explicit 'Use when' and 'Invoke for' clauses plus a dedicated trigger terms list. | 3 / 3 |
Trigger Term Quality | Excellent coverage of natural terms users would say, including both technical and common variations: LoRA, QLoRA, PEFT, finetuning, fine-tuning, adapter tuning, LLM training, model training, custom model, Hugging Face PEFT, OpenAI fine-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, and RLHF are unlikely to conflict with other skills. | 3 / 3 |
Total | 12 / 12 Passed |
Implementation
77%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, actionable skill with a clear multi-step workflow, explicit validation checkpoints, and a complete executable code example covering LoRA, QLoRA, and adapter merging. Its main weaknesses are the missing bundle reference files (which undermines the progressive disclosure design) and minor verbosity in the opening line and constraints section. Overall it would serve Claude well as practical fine-tuning guidance.
Suggestions
Provide the referenced bundle files (references/lora-peft.md, references/dataset-preparation.md, etc.) so the progressive disclosure structure is functional rather than aspirational.
Remove the persona sentence ('Senior ML engineer specializing in...') — it adds no actionable value and wastes tokens.
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | The skill is mostly efficient with well-commented code and useful inline guidance, but includes some unnecessary framing ('Senior ML engineer specializing in...') and the constraints section partially restates what's already implied by the workflow. The code example is long but justified given the complexity of the task. | 2 / 3 |
Actionability | Provides a fully executable, copy-paste-ready LoRA fine-tuning example with real library imports, concrete hyperparameter values with explanatory comments, QLoRA variant, and adapter merging code. The dataset format, training arguments, and deployment steps are all specific and complete. | 3 / 3 |
Workflow Clarity | The 5-step core workflow has explicit validation checkpoints at each stage (dataset validation before training, loss curve monitoring during training, held-out evaluation before deployment). Feedback loops are present — e.g., 'fix all errors before proceeding' and 'plateau or increase signals overfitting' — providing clear error recovery guidance. | 3 / 3 |
Progressive Disclosure | The reference table with 5 topic-specific files and clear 'Load When' triggers is well-structured, but no bundle files were provided, meaning none of those references actually exist. The main SKILL.md also includes a substantial inline code example (~60 lines) that could arguably live in a referenced file, though for a primary working example this is defensible. | 2 / 3 |
Total | 10 / 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
5e8b6b8
- 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.