phase1-method-design

Orchestrate the Architect-Confirmer loop until Phase 1 converges.

Updated Apr 19, 2026
One-click install
npx skills add https://github.com/YurainSoon/research-director --skill phase1-method-design
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: phase1-method-design
Source: https://github.com/YurainSoon/research-director/tree/main/.claude/skills/phase1-method-design
Command: npx skills add https://github.com/YurainSoon/research-director --skill phase1-method-design

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill includes references (resource) components.

What problem does it solve?

Automates the Phase 1 Method Design loop by coordinating the Architect and Confirmer, enforcing round limits, detecting convergence, and coordinating handoff to Phase 2, so human oversight is minimized during routine iterations.

Core Features & Use Cases

  • Orchestrates Round-based workflow between Architect and Confirmer with cap and convergence detection.
  • Determines starting round and manages revision rounds per Case A or Case B scenarios.
  • Handles exit checks and escalation paths for deadlock, and logs outputs for audit.

Quick Start

Start Phase 1 by initiating the Architect → Confirmer loop and monitor rounds until convergence, then hand off to Phase 2

Frequently Asked Questions about phase1-method-design

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I automate research pipeline method design iterations?

Automate research pipeline method design by orchestrating an Architect and Confirmer loop that enforces round limits, detects convergence, and coordinates handoff to Phase 2.

What is the best way to coordinate Architect and Confirmer roles in a research pipeline?

Coordinate Architect and Confirmer roles by running a round-based loop that manages initial runs and revisions, archives design versions, and surfaces Confirmer verdicts until ALL CLEAR.

How does convergence detection work in a method design loop?

Convergence detection in a method design loop works by applying exit checks to Confirmer verdicts, triggering a handoff to Phase 2 once all rounds receive an ALL CLEAR status.

Can I use this method design orchestration without Phase 0 IDEA and EXPERIMENT_SETUP?

This method design orchestration applies only when Phase 0 IDEA and EXPERIMENT_SETUP already exist and Phase 1 has not yet converged, covering initial runs and revisions.

What happens if the Architect and Confirmer reach a deadlock during method design?

If the Architect and Confirmer reach a deadlock during method design, the orchestration handles exit checks and provides escalation paths for the deadlock while logging outputs for audit.

Why does my Phase 1 method design loop keep running without converging?

A Phase 1 method design loop may run without converging if round limits are not enforced properly, requiring escalation paths for deadlock or manual review of archived design versions.