aris-research-refine

Refines vague research directions into concrete method proposals via iterative GPT-5.4 review rounds.

1.1k|116|Updated Feb 26, 2026
One-click install
npx skills add https://github.com/OpenLAIR/dr-claw --skill aris-research-refine
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: aris-research-refine
Source: https://github.com/OpenLAIR/dr-claw/tree/main/skills/aris-research-refine
Command: npx skills add https://github.com/OpenLAIR/dr-claw --skill aris-research-refine

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

Researchers often have a visible problem but a fuzzy technical route, leading to bloated proposals, contribution sprawl, or drift away from the original bottleneck. This Skill turns a vague research direction into a problem-anchored, implementation-ready method plan through structured multi-round external review.

Core Features & Use Cases

  • Problem Anchor Freezing: Extracts an immutable bottom-line problem, constraints, and success conditions that every revision round must preserve, with explicit drift detection.
  • Iterative GPT-5.4 Review Loop: Sends proposals to a Codex-based reviewer across up to 5 rounds, scoring 7 dimensions (problem fidelity, method specificity, contribution quality, frontier leverage, feasibility, validation focus, venue readiness) until the score reaches 9 or the round cap is hit.
  • Checkpoint Recovery: Persists state to REFINE_STATE.json after each phase so interrupted sessions can resume from the last completed phase within 24 hours.
  • Use Case: A PhD student has a rough idea for improving long-context reasoning in LLMs. The skill scans local papers, identifies the technical gap, drafts a focused method proposal with minimal claim-driven experiments, and iteratively refines it against reviewer feedback until it is top-venue ready, then hands off to an experiment-planning skill.

Quick Start

Ask the assistant to refine your research approach by describing your problem and rough idea, for example: refine my approach to reducing hallucination in retrieval-augmented generation using a lightweight verifier.

Frequently Asked Questions about aris-research-refine

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

FAQPage Schema
How do I refine a vague research idea into a concrete method plan?

Provide your problem and rough approach as input. The skill freezes a Problem Anchor, scans local papers for grounding, drafts a focused proposal with method details and minimal experiments, then iterates with an external reviewer until the plan scores 9 out of 10 or reaches 5 rounds.

What does the GPT-5.4 reviewer evaluate in each round?

The reviewer scores seven dimensions: problem fidelity, method specificity, contribution quality, frontier leverage, feasibility, validation focus, and venue readiness. It also flags drift from the original problem, simplification opportunities, and modernization opportunities, returning a READY, REVISE, or RETHINK verdict.

Can a refinement session resume after interruption?

Yes. The skill writes REFINE_STATE.json after each phase with the current phase, round, reviewer thread ID, and last score. If the state is under 24 hours old and marked in_progress, the session resumes from the next phase; older or completed states trigger a fresh start.

Does this skill require the Codex MCP integration?

Yes. External review rounds call the mcp__codex__codex tool with model gpt-5.4 and xhigh reasoning effort, and later rounds reuse the saved threadId via codex-reply. Without Codex MCP access, the review loop cannot execute.

When should I not use iterative research refinement?

Avoid it when you have no defined problem yet, since the skill assumes a visible problem with a fuzzy route. For open-ended ideation, use an idea-generation skill first; for execution-ready experiment roadmaps after refinement, hand off to an experiment-planning skill.