spec-refine-loop

Refine ambiguous draft specs into precise, testable requirements via a gap-driven loop.

Updated May 2, 2026
One-click install
npx skills add https://github.com/Tacuchi/agent-workflow-cli --skill spec-refine-loop
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: spec-refine-loop
Source: https://github.com/Tacuchi/agent-workflow-cli/tree/main/skills/w/loops/spec-refine-loop
Command: npx skills add https://github.com/Tacuchi/agent-workflow-cli --skill spec-refine-loop

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

The spec-refine-loop provides a reusable chassis for refining draft specifications into unambiguous refined specs using a gap-driven, convergent loop that queries humans when needed and can perform autonomous research when appropriate, integrating UI design when required.

Core Features & Use Cases

  • Chassis for refactoring ambiguous specs into precise requirements through iterative refinement.
  • Supports on-demand research sessions and a human-in-the-loop workflow for disambiguation.
  • Integrates with UI design capabilities via a built-in ui-spec when needed.

Quick Start

Invoke the spec-refine-loop by running the /w:spec-refine command on a draft spec to begin the refinement process.

Frequently Asked Questions about spec-refine-loop

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

FAQPage Schema
How do I refine ambiguous draft specifications into precise testable requirements?

You refine ambiguous draft specifications into precise testable requirements by running a gap-driven AI loop that identifies missing details and queries you for disambiguation. The spec refinement process operates directly on your docs/specs drafts.

What is a gap-driven refinement loop for software documentation?

A gap-driven refinement loop for software documentation is an iterative process that analyzes draft specs, identifies ambiguities or missing information, and performs on-demand research or human-in-the-loop validation to converge on unambiguous requirements.

How do I start refining a spec draft using an automated workflow?

You start refining a spec draft by invoking the /w:spec-refine command on your draft specification file. The workflow then initiates the convergent AI loop to begin identifying gaps and generating refined testable requirements.

Does the spec refinement workflow support human-in-the-loop validation and autonomous research?

Yes, the spec refinement workflow supports human-in-the-loop validation by querying you when gaps require clarification, and it can also perform autonomous on-demand research to gather necessary context when appropriate.

Can I integrate UI design specifications into the spec refinement process?

Yes, you can integrate UI design specifications into the spec refinement process. The workflow includes a built-in ui-spec capability to handle UI design requirements when needed during the gap analysis and refinement loop.

Where are refined specs and progress checkpoints persisted during the workflow?

Refined specs are persisted to docs/specs/NNN-spec-refined.md, while progress checkpoints and backlog artifacts are saved in .workflow/sessions. This ensures continuity and resilience across refinement sessions.