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.
38
37%
Does it follow best practices?
Impact
—
No eval scenarios have been run
Passed
No known issues
Optimize this skill with Tessl
npx tessl skill review --optimize ./plugins/ai-ml/dataset-splitter/skills/splitting-datasets/SKILL.mdQuality
Discovery
75%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 adequately covers the what and when, earning strong completeness and distinctiveness scores. However, it lacks depth in specific capabilities (e.g., stratification, ratio configuration, supported formats) and the final sentence 'Trigger with relevant phrases based on skill purpose' is vague filler that adds no value. Trigger term coverage could be broader to capture more natural user phrasings.
Suggestions
Add more concrete actions beyond basic splitting, such as 'configure split ratios, apply stratified sampling, handle class imbalance, export partitioned datasets'.
Expand trigger terms to include common variations like 'holdout set', 'stratified split', 'cross-validation folds', 'train/val/test ratio'.
Remove the vague final sentence 'Trigger with relevant phrases based on skill purpose' as it provides no useful information for skill selection.
| Dimension | Reasoning | Score |
|---|---|---|
Specificity | Names the domain (ML dataset splitting) and the core action (process split datasets into train/val/test sets), but doesn't list multiple concrete actions beyond the single splitting operation—no mention of stratification, shuffling, ratio configuration, or output formats. | 2 / 3 |
Completeness | Explicitly answers both 'what' (process split datasets into training, validation, and testing sets for ML model development) and 'when' (use when requesting 'split dataset', 'train-test split', or 'data partitioning') with clear trigger guidance. | 3 / 3 |
Trigger Term Quality | Includes some natural trigger terms like 'split dataset', 'train-test split', and 'data partitioning', but misses common variations users would say such as 'holdout set', 'cross-validation', 'stratified split', 'train/val/test', or file format references like CSV/parquet. | 2 / 3 |
Distinctiveness Conflict Risk | The description targets a clear niche—ML dataset splitting into train/val/test sets—with specific trigger terms that are unlikely to conflict with general data processing or other ML skills. | 3 / 3 |
Total | 10 / 12 Passed |
Implementation
0%Reviews the quality of instructions and guidance provided to agents. Good implementation is clear, handles edge cases, and produces reliable results.
This skill is almost entirely boilerplate with no actionable content. It lacks any executable code, concrete commands, or specific library usage (e.g., sklearn.model_selection.train_test_split). The majority of sections contain generic filler text that could apply to any skill, providing no unique value for the dataset splitting task.
Suggestions
Replace the abstract 'How It Works' and 'Examples' sections with actual executable Python code using sklearn's train_test_split, including stratification and random seed parameters.
Remove all boilerplate sections (Integration, Prerequisites, Instructions, Output, Error Handling, Resources) that contain only generic placeholder text and add no task-specific value.
Add a concrete validation step, such as printing the shape of each split and verifying class distributions, to ensure the split was performed correctly.
Include a concise code example showing configurable ratios, stratification on a target column, and saving outputs to CSV files—all in under 30 lines.
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | Extremely verbose with extensive padding. Sections like 'Overview', 'When to Use This Skill', 'Integration', 'Prerequisites', 'Instructions', 'Output', 'Error Handling', and 'Resources' are all filler that explain nothing Claude doesn't already know. The 'How It Works' section describes obvious steps (analyze request, generate code, execute). Most of the content is boilerplate that adds no actionable value. | 1 / 3 |
Actionability | No executable code anywhere. The examples describe what the skill 'will do' in abstract terms rather than providing actual Python code using sklearn's train_test_split or pandas. Instructions like 'Invoke this skill when the trigger conditions are met' and 'Provide necessary context and parameters' are completely vague. There is nothing copy-paste ready or concrete. | 1 / 3 |
Workflow Clarity | The workflow steps ('Analyze Request', 'Generate Code', 'Execute Splitting') are abstract meta-descriptions, not actual operational steps. There are no validation checkpoints—no verification that splits sum to 100%, no check for data leakage, no confirmation of output file integrity. The 'Instructions' section is generic boilerplate with no real sequence. | 1 / 3 |
Progressive Disclosure | The content is a monolithic wall of boilerplate text with no bundle files to reference. The 'Resources' section points to nothing specific ('Project documentation', 'Related skills and commands'). There is no meaningful structure—just many shallow sections that repeat vague platitudes rather than organizing content for discovery. | 1 / 3 |
Total | 4 / 12 Passed |
Validation
81%Checks the skill against the spec for correct structure and formatting. All validation checks must pass before discovery and implementation can be scored.
Validation — 9 / 11 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 | 9 / 11 Passed | |
- Commit
6e9558f
- 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.