Where are the specification files stored?

All generated specifications are saved to a local /specs directory within your project path, following standardized naming conventions for easy organization.

What happens if a spec for a specific issue already exists?

The skill is idempotent; it will skip creation and return the existing spec info unless you use the 'force' flag to generate an additional version.

How does the template auto-detection work?

The skill analyzes the merged context of your conversation and linked GitHub issue to classify the work type as a feature, bug, infrastructure, or API task before selecting the appropriate template.

Can I use Spec Generator without a GitHub connection?

Yes. If no GitHub repository or issue ID is detected, the skill gracefully degrades to creating a standalone specification using a timestamped naming convention.

How does Spec Generator capture full conversation context?

It is invoked directly via a specialized command that bypasses the agent layer, ensuring all planning discussions and context are preserved in the generated specification.

Spec Generator

Name: Spec Generator
Author: fractary

$fractary$ byfractary

•

Productivity & Workflow

Automatically generates detailed implementation specifications by merging conversation context with GitHub issue data.

The Spec Generator skill streamlines the transition from planning to execution by capturing full conversation context and merging it with GitHub issue data to produce standardized implementation documents. It intelligently classifies work types—such as features, bugs, or infrastructure—to apply the most relevant template, ensuring all technical requirements and architectural decisions are documented in a centralized directory. By bypassing the standard agent layer during invocation, it preserves the nuances of complex planning discussions that might otherwise be lost, providing a reliable source of truth for development teams.

Key Features

01Intelligent work classification for Feature, Bug, API, and Infrastructure templates

02Standardized file naming conventions and frontmatter metadata management

03Context-aware spec generation directly from AI chat history

042 GitHub stars

05Automatic GitHub issue data enrichment and bi-directional linking

06Idempotent creation logic with force-override support for iterative planning

Use Cases

01Maintaining a versioned, local history of implementation decisions within a project's specs directory

02Converting a complex architectural brainstorming session into a structured technical design document

03Syncing GitHub feature requests with local implementation specs to ensure requirement coverage

What are Skills?·How to Install

Install with 🐟 Skill.Fish

npx skillfish add fractary/claude-plugins spec-generator

For use in Claude.ai and ChatGPT

Download Skill