Where are the extracted requirements stored?

Results are saved as structured YAML files within a local .requirements/ directory, organized by the domain name you specify.

How does it handle duplicate requirements across multiple files?

The skill includes a dedicated deduplication step that identifies overlapping requirements and recommends the most specific version for consolidation.

Can I use this for competitor analysis?

Yes, by using the --type competitor flag and providing a URL, the skill can crawl external sites and convert feature lists into requirement candidates.

What is the difference between Guided and Full-auto modes?

Guided mode prompts you to confirm every requirement candidate the AI finds, while Full-auto processes the entire document and provides a summary report at the end.

What file formats does the extract skill support?

The skill supports PDF, Markdown (.md), Plain Text (.txt), Word (.docx), and live web URLs via the WebFetch tool.

AI Requirement Extractor

Name: AI Requirement Extractor
Author: melodic-software

bymelodic-software

•

Learning & Documentation

Extracts and categorizes software requirements from documents and URLs into structured, canonical YAML formats.

The AI Requirement Extractor is a specialized tool designed to automate the process of requirement mining and elicitation. By scanning various sources—including PDFs, Markdown files, Word documents, and live websites—it identifies functional and non-functional requirements, constraints, and assumptions with high precision. The skill utilizes a document-miner agent to handle complex layouts and provides multiple autonomy modes to balance speed with human oversight, ensuring that project specifications are comprehensive, deduplicated, and ready for development.

Key Features

01Three autonomy modes: Guided, Semi-auto, and Full-auto for flexible control

0238 GitHub stars

03Automated categorization into functional, non-functional, and constraint types

04Multi-format support for PDF, Word, Markdown, and live web URLs

05Cross-document deduplication to ensure a single source of truth

06Source attribution with page and section tracking for every requirement

Use Cases

01Conducting competitor feature analysis by extracting requirements from public documentation

02Consolidating multiple meeting transcripts and rough notes into a formal project scope

03Transforming legacy PDF specifications into structured data for agile backlogs

What are Skills?·How to Install

Install with 🐟 Skill.Fish

npx skillfish add melodic-software/claude-code-plugins extract

For use in Claude.ai and ChatGPT

Download Skill