meta-loop-orchestrator

---

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

Coordinate iterative improvement loops (e.g., RICE, dogfooding, COV) to refine deliverables while preventing endless iteration or confidence drift.

Trigger Conditions

• Positive: recursive improvement, self-critique loops, A/B iteration planning, convergence checks, dogfooding cycles.
• Negative: single-pass edits, straightforward prompt cleanups (route to prompt-architect), or new skill weaving (route to skill-forge).

Guardrails

• Skill-Forge structure-first: keep SKILL.md, examples/, tests/ updated; add resources//references/ or document gaps.
• Prompt-Architect hygiene: extract HARD/SOFT/INFERRED goals per iteration, define stop criteria, and state ceilings for confidence.
• Loop safety: set iteration caps, convergence thresholds, and rollback checkpoints; enforce registry-only agents and hook latencies.
• Adversarial validation: challenge assumptions every loop, run boundary tests, and record evidence plus deltas.
• MCP tagging: log loops with WHO=meta-loop-orchestrator-{session} and WHY=skill-execution.

Execution Playbook

1. Intent & target: define improvement goal, metrics, and stop criteria; confirm inferred needs.
2. Loop design: select frameworks (COV, dogfooding), assign roles, and timebox iterations.
3. Execution: run iteration, collect evidence, compute delta; keep artifacts versioned.
4. Adversarial check: probe edge cases, challenge assumptions, and update risk register.
5. Convergence decision: compare delta to threshold; decide continue, pivot, or stop.
6. Delivery: summarize iterations, evidence, residual risks, and confidence ceiling.

Output Format

• Improvement goal, metrics, and stop rule.
• Iteration log (changes, evidence, deltas).
• Risk/assumption register and next steps.
• Confidence: X.XX (ceiling: TYPE Y.YY) - rationale.

Validation Checklist

• Structure-first assets present or ticketed; examples/tests reflect current best version.
• Iteration caps and stop rules respected; registry and hooks validated.
• Adversarial/COV evidence stored with MCP tags; confidence ceiling declared; English-only output.

Completion Definition

Loop concludes when delta falls below threshold or stop rule triggers, evidence is stored, risks are owned, and a ready-to-use version is documented with confidence ceiling.

Confidence: 0.70 (ceiling: inference 0.70) - Meta-loop SOP aligned to skill-forge scaffolding and prompt-architect clarity with explicit convergence controls.