gsd-spec-phase

Run structured Socratic questioning with ambiguity scoring to write SPEC.md.

1|Updated Sep 6, 2015
One-click install
npx skills add https://github.com/deerawan/dotfiles --skill gsd-spec-phase-deerawan
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: gsd-spec-phase
Source: https://github.com/deerawan/dotfiles/tree/main/claude/skills/gsd-spec-phase
Command: npx skills add https://github.com/deerawan/dotfiles --skill gsd-spec-phase-deerawan

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

It removes ambiguity about what a workflow phase must deliver by running a structured Socratic questioning loop and generating a falsifiable SPEC.md before the next phase proceeds.

Core Features & Use Cases

  • Socratic clarification with ambiguity scoring: Runs up to 6 rounds of rotating-perspective questions and quantifies ambiguity across four weighted dimensions after each round.
  • Spec-gated workflow progression: Only writes SPEC.md when ambiguity is low enough (≤ 0.20) and all minimum dimension thresholds are met, preventing “guess-based” downstream work.
  • Codebase-aware requirements: Scans the current repository state before asking questions so the SPEC reflects real constraints and facts.
  • Automatic handoff to discuss-phase: Commits SPEC.md so the subsequent discuss-phase can load it on the next run and focus on the “how”.

Quick Start

Run gsd-spec-phase with the phase name you want to specify (for example, “plan-phase”) to generate a padded SPEC.md that locks down the phase’s deliverables and acceptance criteria.

Frequently Asked Questions about gsd-spec-phase

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

FAQPage Schema
How do I turn phase intent into an unambiguous specification before starting development?

To turn phase intent into an unambiguous specification, you run a structured Socratic questioning loop that quantifies ambiguity across four dimensions. It generates a falsifiable SPEC.md that locks down deliverables and acceptance criteria before development proceeds.

What is Socratic questioning with ambiguity scoring for requirements gathering?

Socratic questioning with ambiguity scoring is a structured interview loop that asks up to six rounds of rotating-perspective questions to clarify deliverables. It quantifies ambiguity after each round to ensure requirements meet a strict threshold before workflow progression.

How do I prevent guess-based downstream work in my workflow planning?

You prevent guess-based downstream work by applying spec-gated workflow progression that only writes a committed SPEC.md when ambiguity scoring drops to 0.20 or below. This ensures all minimum dimension thresholds are met before the next phase begins.

Does the specification process scan the current codebase before generating requirements?

Yes, the specification process scans the current repository state before asking questions to ensure the generated SPEC.md reflects real constraints and facts. This codebase-aware approach grounds the Socratic interview in actual project conditions.

What happens after the SPEC.md is committed in the workflow?

After the SPEC.md is committed, the workflow automatically hands off to the discuss-phase. The subsequent discuss-phase loads the committed specification on the next run to focus entirely on determining how to execute the locked deliverables.

When should I use automated specification generation in my workflow?

You should use automated specification generation after the codebase is first scouted and context files are loaded, but before the discuss-phase begins. It is required when you need to define phase deliverables with quantitative ambiguity scoring before execution.