splitting-datasets

Process split datasets into training, validation, and testing sets for ML model development. Use when requesting "split dataset", "train-test split", or "data partitioning". Trigger with relevant phrases based on skill purpose.

Quality

55%

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/ai-ml/dataset-splitter/skills/splitting-datasets/SKILL.md

Quality

Content

20%

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

The content is mostly generic filler with no executable code, no named library, and no validation steps, despite bundled assets and (todo) scripts being available. It describes the skill in the abstract instead of giving Claude concrete, copy-paste-ready guidance.

Suggestions

Replace the 'How It Works', 'Instructions', 'Output', and 'Resources' filler with a concrete executable Python example using a named library (e.g., scikit-learn train_test_split / StratifiedShuffleSplit) that produces train/validation/test CSVs.

Reference the bundled assets directly — link to assets/split_data_config.yaml, assets/example_dataset.csv, and assets/dataset_schema.json, and invoke scripts/split_data.py instead of describing code generation in the abstract.

Add explicit validation checkpoints to the workflow (verify split ratios sum correctly, check for data leakage, confirm class distribution after stratified splitting) and remove the placeholder 'Resources' entries.

Dimension	Reasoning	Score
Conciseness	The body is padded with filler Claude already knows ("The skill produces structured output relevant to the task", generic best-practice advice, placeholder "Resources" entries) and restates the title in multiple sections rather than adding new information.	1 / 3
Actionability	No executable code or commands appear; it only says it "generates Python code utilizing standard ML libraries" without naming a library, and the examples describe outcomes rather than instructing, matching the level-1 anchor.	1 / 3
Workflow Clarity	A three-step sequence (Analyze Request, Generate Code, Execute Splitting) is present, but for a batch data operation there are no validation or verification checkpoints, which caps the score at 2 per the rubric guideline.	2 / 3
Progressive Disclosure	Section headers provide some structure, but the body never signals or links to the actual bundle files (assets/dataset_schema.json, assets/split_data_config.yaml, scripts) and the "Resources" section is placeholder text rather than real navigation.	2 / 3
	Total	6 / 12 Passed

Description

90%

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 answers both what the skill does and when to use it with concrete, natural trigger phrases and a clear niche. Its only weakness is the single awkward "Process split datasets" action phrasing and a trailing vague sentence ("Trigger with relevant phrases based on skill purpose.") that adds nothing.

Dimension	Reasoning	Score
Specificity	It names the ML domain and one core action ("Process split datasets into training, validation, and testing sets"), but describes a single splitting operation rather than listing multiple distinct concrete actions like the level-3 anchor.	2 / 3
Completeness	It explicitly states what the skill does and includes a "Use when requesting ..." clause with concrete triggers, satisfying both the what and the when at the level-3 anchor.	3 / 3
Trigger Term Quality	It supplies three natural phrases a user would actually say ("split dataset", "train-test split", "data partitioning"), giving good coverage of the common ways this task is requested.	3 / 3
Distinctiveness Conflict Risk	The ML dataset-splitting niche with specific triggers is clearly distinguishable and unlikely to fire for unrelated skills.	3 / 3
	Total	11 / 12 Passed

Validation

87%

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 — 14 / 16 Passed

Validation for skill structure

Criteria	Description	Result
allowed_tools_field	'allowed-tools' contains unusual tool name(s)	Warning
frontmatter_unknown_keys	Unknown frontmatter key(s) found; consider removing or moving to metadata	Warning

	Total	14 / 16 Passed

Repository: jeremylongshore/claude-code-plugins-plus-skills
Path: plugins/ai-ml/dataset-splitter/skills/splitting-datasets/SKILL.md
Commit: 15940c5

Reviewed: 4 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.