Where are the generated specification files saved?

Specifications are automatically generated as Markdown files and saved in the .aiwork/specs/ directory with a standardized naming convention.

Can I customize the format of the output spec?

The skill uses a template located at assets/spec-template.md, which can be modified to fit your team's specific documentation requirements.

How does the Spec skill gather initial project context?

The skill can read triage files from the .aiwork/triage/ directory, fetch ticket metadata from Azure DevOps, or continue directly from a previous triage session.

What kind of questions does the skill ask the user?

It focuses on four main categories: Architecture Decisions (layers, API styles), Dependencies (blocking tasks), Implementation Approach (migration strategies), and Process (PR strategy, testing).

Does it identify which files will be affected by a change?

Yes, it uses an exploration agent to find current component structures, props, and all consumer files that import the components in scope.

DevFlow Spec Generator

Name: DevFlow Spec Generator
Author: lttr

bylttr

•

Productividad y Flujo de Trabajo

Transforms triage data and user decisions into detailed implementation specifications for streamlined development.

DevFlow Spec is a specialized Claude Code skill designed to bridge the gap between initial task triage and actual coding. It automates the creation of comprehensive implementation specs by analyzing triage files or Azure DevOps tickets, resolving architectural ambiguities through structured questioning, and performing technical discovery. By producing standardized Markdown specifications that include dependency maps, consumer file lists, and acceptance criteria, it ensures developers have a clear, verified roadmap before writing a single line of code.

Características Principales

01Structured 'AskUserQuestion' workflow for resolving architecture and dependency blockers

02Automated spec generation from triage reports or Azure DevOps ticket IDs

03Automated technical discovery of component interfaces and consumer files

04Standardized Markdown output using professional implementation templates

05Dependency graph generation to establish explicit execution order

061 GitHub stars

Casos de Uso

01Formalizing architectural decisions and migration strategies for complex refactorings

02Converting high-level feature requests into actionable, step-by-step technical plans

03Ensuring consistent documentation and testing standards across a distributed engineering team

What are Skills?·How to Install

Install with 🐟 Skill.Fish

npx skillfish add lttr/claude-marketplace spec

For use in Claude.ai and ChatGPT

Download Skill