Create community playbook: ML/Data Science

[VictorVVedtion]

Description

Create a community playbook in playbooks/ml-ops/ that adapts selfmodel for machine learning and data science projects.

Suggested customizations

• Quality gates: Add "Data Integrity" dimension (detect train/test leakage, check for data drift)
• Dispatch rules: Route data pipeline tasks to Codex, model architecture to Opus
• Evaluator prompt: Check for reproducibility (random seeds, deterministic operations)
• Sprint template: Add "Dataset" and "Metrics" fields

Files to create

playbooks/ml-ops/
  README.md
  quality-gates.md
  dispatch-rules.md
  evaluator-prompt.md

See playbooks/README.md for the playbook specification.

Acceptance criteria

• Playbook directory with README and at least 2 rule files
• README explains the ML-specific adaptations
• Rules follow Iron Rules (no mock data, etc.)