tiemaideradicate/r04-alirezarezvani-claude-code-skill-factory-datascience
ๅผๅๅทฅๅ ท๐ค Data Science & AI/ML skill suite derived from alirezarezvani/claude-code-skill-factory.
ๆฆ่ง
Source focus: skill scaffolding, code agents, slash command generation Data pipelines, model training, evaluation, MLOps and analytical reporting. This collection provides and , all with a consistent structured-output UI so you always know exactly where you are and what to do next. All commands display structured output with: - โ real-time step tracking - โ sorted by severity (๐ด๐ ๐ก๐ข) - โ quick wins โ medium-term โ strategic - โ at-a-glance metrics after each command Every command follows this 5-step structure: This suite is derived from which focuses on: skill scaffolding, code agents, slash command generation. Improvements in this adaptation: - Domain-specific command vocabulary for Data Science & AI/ML - Enhanced structured output with visual progress tracking - Prioritised action plans with time estimates - Workflow orchestration for end-to-end processes - Consistent UI conventions across all commands
README
๐ค Data Science & AI/ML Skills Suite
Derived from alirezarezvani/claude-code-skill-factory
Adaptation of
alirezarezvani/claude-code-skill-factoryfor Data Science & AI/ML use cases. Source focus: skill scaffolding, code agents, slash command generation
What This Skill Suite Does
Data pipelines, model training, evaluation, MLOps and analytical reporting.
This collection provides 10 specialised commands and 5 multi-step workflows, all with a consistent structured-output UI so you always know exactly where you are and what to do next.
Quick Install
# Clone this skill
cp -r . ~/.claude/skills/r00-alirezarezvani-claude-code-skill-factory--datascience/
# Register in Claude Code
# In a Claude Code session:
/read ~/.claude/skills/r00-alirezarezvani-claude-code-skill-factory--datascience/SKILL.md
Commands
| Command | Description |
|---|---|
/data-profiling |
Automated EDA report: distributions, nulls, outliers, correlations and drift |
/feature-engineer |
Feature importance analysis with SHAP values and automated encoding recipes |
/model-evaluate |
Model performance dashboard: ROC, PR curves, confusion matrix and bias check |
/pipeline-scaffold |
Modular ML pipeline scaffold with versioning, logging and registry hooks |
/ab-test-design |
Statistical A/B test design: sample size, power, MDE and sequential testing |
/sql-optimize |
Query plan analysis, index recommendations and cost estimation |
/dashboard-spec |
BI dashboard specification from KPI list with chart types and data sources |
/data-contract |
Schema validation, SLA definition and data quality contract generation |
/llm-eval |
LLM output evaluation harness: hallucination rate, faithfulness and latency |
/anomaly-detect |
Time-series anomaly detection with root-cause attribution and alert tuning |
Usage:
/data-profiling
/feature-engineer --scope full --output md
Workflows (Multi-step)
| Workflow | Description |
|---|---|
ml-project-init |
End-to-end ML project: EDA โ baseline โ feature engineering โ model โ deploy |
data-migration |
Data warehouse migration: audit โ schema map โ ETL โ validation โ cutover |
reporting-pipeline |
Automated reporting pipeline: source โ transform โ validate โ visualise โ deliver |
model-retraining |
Scheduled model retraining: drift detect โ retrain โ shadow โ promote โ monitor |
analytics-sprint |
2-week analytics sprint: question โ data โ analysis โ insight โ recommendation |
Usage:
/workflows:ml-project-init --scope full
UI Design
All commands display structured output with:
- Progress panels โ real-time step tracking
- Findings tables โ sorted by severity (๐ด๐ ๐ก๐ข)
- Action checklists โ quick wins โ medium-term โ strategic
- Summary cards โ at-a-glance metrics after each command
Progress Display Example
โโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโ
โ ML Pipeline โ churn_prediction_v3 โ
โ โโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโฃ
โ Data ingestion โ 1.2M rows loaded โ
โ Profiling โ 12 features, 3 issues found โ
โ Feature eng. โ 47 features created โ
โ Training โณ Epoch 18/50 [โโโโโโโโโโ] โ
โ Evaluation โ Pending โ
โ Registry push โ Pending โ
โโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโ
DATA QUALITY ISSUES
โ customer_age โ 847 nulls (0.07%) โ Impute with median
โ last_purchase โ 12 future dates โ Clip to today
โ revenue โ 3ฯ outliers: 214 โ Review before drop
Interaction Pattern
Every command follows this 5-step structure:
โ Scope Confirmation โ verify target and options with user
โก Live Analysis โ progress bar while working
โข Findings Table โ structured results sorted by impact
โฃ Action Plan โ prioritised, time-boxed recommendations
โค Next Steps โ suggested follow-up commands
Source Repository
This suite is derived from alirezarezvani/claude-code-skill-factory which focuses on: skill scaffolding, code agents, slash command generation.
Improvements in this adaptation:
- Domain-specific command vocabulary for Data Science & AI/ML
- Enhanced structured output with visual progress tracking
- Prioritised action plans with time estimates
- Workflow orchestration for end-to-end processes
- Consistent UI conventions across all commands
License
MIT โ free to use, modify and distribute.
ๆจ่ๅทฅๅ ท
ๆขไธไธชๅ ณ้ฎ่ฏ๏ผๆ่ ็งป้ค็ญ้ๆกไปถใ
ๅฎ่ฃ
npx skillfish add tiemaideradicate/r04-alirezarezvani-claude-code-skill-factory-datascience