ml

---

LIBRARY-FIRST PROTOCOL (MANDATORY)

Before writing ANY code, you MUST check:

Step 1: Library Catalog

• Location: .claude/library/catalog.json
• If match >70%: REUSE or ADAPT

Step 2: Patterns Guide

• Location: .claude/docs/inventories/LIBRARY-PATTERNS-GUIDE.md
• If pattern exists: FOLLOW documented approach

Step 3: Existing Projects

• Location: D:\Projects\*
• If found: EXTRACT and adapt

Decision Matrix

Match	Action
Library >90%	REUSE directly
Library 70-90%	ADAPT minimally
Pattern exists	FOLLOW pattern
In project	EXTRACT
No match	BUILD (add to library after)

---

STANDARD OPERATING PROCEDURE

Purpose

Scope and deliver ML prototypes or lightweight solutions, then hand off to the right specialist (ml-expert, ml-training-debugger) when depth or incident response is required.

Triggers

• Positive: Early ML ideation, quick prototypes, feature feasibility, light model improvements.
• Negative: Complex training/debugging (route to ml-expert or ml-training-debugger) or pure prompt design (use prompt-architect).

Guardrails

• Structure-first: keep SKILL.md, readme, examples/, tests/, resources/ up to date; create missing docs before execution.
• Constraint clarity: HARD/SOFT/INFERRED requirements captured; ambiguous items confirmed.
• Validation: baseline vs simple ablation; sanity checks on data splits and metrics.
• Confidence ceilings enforced (inference/report 0.70; research 0.85; observation/definition 0.95).
• Ethical/compliance: avoid biased data, note privacy/security constraints.

Execution Phases

1. Intake & Scoping
   • Define objective, target metric, timeline, and constraints.
   • Identify data sources, size, and quality risks.
2. Design & Plan
   • Select a simple, reliable baseline; outline minimal pipeline and evaluation plan.
   • Choose handoff target if deeper expertise will be needed.
3. Prototype
   • Implement baseline with reproducible configs and logging.
   • Add small improvements (feature engineering, regularization, lightweight tuning).
4. Validate
   • Evaluate on validation split; report metrics with variance.
   • Run basic robustness checks (class imbalance, leakage, overfitting signs).
5. Handoff/Delivery
   • Provide code, configs, data notes, and next-step recommendations.
   • Route to ml-expert or ml-training-debugger with context if further work is required.

Output Format

• Request summary and constraints.
• Baseline choice, experiments run, and metrics.
• Risks, limitations, and recommended next steps/handoff.
• Confidence with ceiling.

Validation Checklist

• Constraints captured and confirmed.
• Data split documented; leakage check done.
• Baseline + at least one improvement executed.
• Metrics reported with source split.
• Confidence ceiling stated.

VCL COMPLIANCE APPENDIX (Internal)

[[HON:teineigo]] [[MOR:root:M-L]] [[COM:Genel+ML]] [[CLS:ge_skill]] [[EVD:-DI<gozlem>]] [[ASP:nesov.]] [[SPC:path:/skills/specialists/ml]]

[[HON:teineigo]] [[MOR:root:E-P-S]] [[COM:Epistemik+Tavan]] [[CLS:ge_rule]] [[EVD:-DI<gozlem>]] [[ASP:nesov.]] [[SPC:coord:EVD-CONF]]

Confidence: 0.72 (ceiling: inference 0.70) - SOP rewritten with skill-forge structure and prompt-architect constraint discipline.