create-spec

Convert feature ideas or issues into Ralph execution specs and GitHub issues.

1|4|Updated Oct 27, 2014
One-click install
npx skills add https://github.com/rjernst/dotfiles --skill create-spec-rjernst
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: create-spec
Source: https://github.com/rjernst/dotfiles/tree/main/claude/skills/create-spec
Command: npx skills add https://github.com/rjernst/dotfiles --skill create-spec-rjernst

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

It turns a feature idea or an existing issue into a precise, self-contained execution spec that Ralph can run, reducing back-and-forth and missing requirements.

Core Features & Use Cases

  • Interactive spec drafting: Guides the user through feature name, scope, constraints, acceptance criteria, and a checkable task breakdown without writing code.
  • Issue-aware planning: Detects numeric/URL/natural-language input, looks up or searches issues in the user’s fork first, and can base the spec on a source issue’s goal.
  • GitHub issue creation for execution: Creates a labeled GitHub issue in the user’s fork with the generated spec body so Ralph can execute via its dockerized coding loop.

Quick Start

Use the create-spec skill by invoking /create-spec with your feature idea or an issue number, and it will generate a reviewable spec issue body ready for Ralph execution.

Frequently Asked Questions about create-spec

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

FAQPage Schema
How do I generate an executable spec from a GitHub issue for automated planning?▼

To generate an executable spec from a GitHub issue, you can use an automation tool to convert the issue into a self-contained, checkable execution spec with acceptance criteria and a templated task breakdown. This reduces missing requirements and back-and-forth during automated planning.

How do I create a GitHub issue with a feature spec in my fork?▼

You can create a GitHub issue with a feature spec in your fork by using a CLI tool that resolves the repository via git remote get-url and executes gh issue create with an approved body file. This populates your fork with a labeled issue ready for automated execution.

Can I search for existing issues in natural language to draft a feature spec?▼

Yes, you can search for existing issues using natural language to draft a feature spec. The planning agent detects natural-language input, searches issues in your fork, and bases the generated execution spec on the source issue's goal and context.

Does spec generation work with interactive workflows without prior code?▼

Yes, spec generation works with interactive workflows without prior code by conducting context-based interviews. It guides you through feature name, scope, constraints, and acceptance criteria to produce a reviewable spec issue body without writing code.

What is included in an executable feature spec for a planning agent?▼

An executable feature spec for a planning agent includes a self-contained, checkable task breakdown with frontmatter branch, base, and depends fields as needed. It also contains acceptance criteria and constraints to ensure the automation loop can run.

Do I need a GitHub fork to create an executable spec issue?▼

Yes, you need a GitHub fork to create an executable spec issue because the tool resolves the target repository using git remote get-url and creates the labeled issue using gh issue create --repo origin-repo within that fork.