ml-model-training

Train ML models with scikit-learn, PyTorch, TensorFlow. Use for classification/regression, neural networks, hyperparameter tuning, or encountering overfitting, underfitting, convergence issues.

Quality

70%

Does it follow best practices?

Run evals on this skill

Adds up to 20 points to the overall score

View guide

Securityby

Passed

No findings from the security scan

Fix and improve this skill with Tessl

tessl review fix ./plugins/ml-model-training/skills/ml-model-training/SKILL.md

Quality

Content

60%Weight 40%Scale 1-5

Reviews the quality of instructions and guidance provided to agents. Good implementation is clear, handles edge cases, and produces reliable results.

The content is a well-organized, actionable overview that leverages real reference files for depth, scoring well on actionability and progressive disclosure. It is held back by duplicated snippets and restated best-practice concepts that hurt conciseness, and by the absence of explicit validation/feedback checkpoints in the training workflow, which caps workflow clarity at 3.

Suggestions

De-duplicate the StandardScaler snippet (shown in both 'Data Preparation' and 'Data Leakage') by referencing the first occurrence, and trim the 'Known Issues Prevention' explanations of concepts Claude already knows to improve conciseness.

Add explicit validation checkpoints and a feedback loop to the workflow (e.g., 'After each epoch, evaluate on val; if val_loss rises for N epochs, apply regularization/early stopping and retrain') to lift workflow clarity above 3.

Define the missing PyTorch tensor inputs (X_train_tensor/y_train_tensor) or note the required conversion from the numpy arrays so the training loop is fully copy-paste ready.

Dimension	Reasoning	Score
Conciseness	The body is mostly efficient with executable code blocks and tight tables, but it repeats material already shown inline (the data-preparation/scaling snippet reappears verbatim in the 'Data Leakage' section) and the 'Known Issues Prevention' section restates best-practice concepts Claude already knows, adding padding.	3 / 5
Actionability	Most guidance is concrete and copy-paste ready (sklearn, PyTorch, and Keras snippets with real imports and parameters), with only minor gaps such as the PyTorch loop referencing undefined tensors (X_train_tensor, y_train_tensor) without showing their creation.	4 / 5
Workflow Clarity	A clear five-step workflow is stated up front, but training is a batch/iterative operation and the body lacks explicit validation checkpoints or feedback loops (e.g., 'evaluate on val, if overfitting then add regularization, re-evaluate'); per the rubric's destructive/batch cap this caps the score at 3.	3 / 5
Progressive Disclosure	Structure is good: the SKILL.md is an overview with well-signaled one-level-deep references to real files (references/pytorch-training.md, references/tensorflow-keras.md) and a dedicated 'When to Load References' section, though some advanced material (full callbacks, checkpointing) is described inline rather than fully delegated to the references.	4 / 5
	Total	14 / 20 Passed

Description

81%Weight 40%Scale 1-5

Based on the skill's description, can an agent find and select it at the right time? Clear, specific descriptions lead to better discovery.

The description is strong: it names concrete frameworks, lists natural trigger phrases including diagnostic terms like overfitting and convergence issues, and provides an explicit 'Use for' clause covering both what and when. It falls just short of top marks because the action verb ('Train') is generic rather than enumerating multiple distinct operations, and the 'when' guidance is phrased as 'Use for' rather than a fully explicit 'Use when'.

Suggestions

Replace the single generic verb 'Train' with several distinct concrete actions (e.g., 'Train, fine-tune, and evaluate ML models...') to lift specificity toward a 5.

Reframe the trigger clause as an explicit 'Use when...' sentence (e.g., 'Use when the user asks to train or tune a model, or reports overfitting, underfitting, or convergence issues') to make completeness unambiguous.

Dimension	Reasoning	Score
Specificity	Names the domain ('Train ML models with scikit-learn, PyTorch, TensorFlow') and several concrete task areas (classification/regression, neural networks, hyperparameter tuning), but the actions themselves are generic ('Train') rather than a varied list of specific operations, leaving minor coverage gaps.	4 / 5
Completeness	Clearly answers 'what' (train ML models with three frameworks) and 'when' via the 'Use for...' clause listing trigger scenarios, but the 'when' phrasing ('Use for') is slightly less explicit than a canonical 'Use when...' formulation and could be more specific.	4 / 5
Trigger Term Quality	Comprehensive coverage of natural user terms including the framework names (scikit-learn, PyTorch, TensorFlow) and common problem phrases users would say ('overfitting, underfitting, convergence issues'), plus 'hyperparameter tuning' and 'classification/regression'.	5 / 5
Distinctiveness Conflict Risk	The framework-specific triggers (PyTorch, TensorFlow, scikit-learn) carve out a distinct niche with minimal overlap risk, though the broad 'ML model training' framing could marginally overlap with adjacent data-science skills.	4 / 5
	Total	17 / 20 Passed

Validation

93%

Warnings & errors only

Checks the skill against the spec for correct structure and formatting. All validation checks must pass before discovery and implementation can be scored.

Validation — 15 / 16 Passed

Validation for skill structure

Criteria	Description	Result
metadata_version	'metadata.version' is missing	Warning

	Total	15 / 16 Passed

Repository: secondsky/claude-skills
Path: plugins/ml-model-training/skills/ml-model-training/SKILL.md
Commit: 84802a4

Reviewed: 3 days ago

Table of Contents

Discovery Implementation Validation

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.