loopeng-extract

Extracts requirements from specifications into numbered checklist ledgers with traceability matrices.

Updated Jun 24, 2026
One-click install
npx skills add https://github.com/Hakkadaikon/hymme --skill loopeng-extract-hakkadaikon
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: loopeng-extract
Source: https://github.com/Hakkadaikon/hymme/tree/main/skills/loopeng-extract
Command: npx skills add https://github.com/Hakkadaikon/hymme --skill loopeng-extract-hakkadaikon

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill includes references (resource) and assets (resource) components.

What problem does it solve? Formal verification tools like TLA+ and Lean rigorously check the requirements they are given, but they cannot guarantee that the extraction of requirements from the original specification was complete. This Skill closes that gap by enforcing an exhaustive, auditable extraction step before any formalization begins, so missing requirements surface at the entry point instead of at the exit. ## Core Features & Use Cases - Exhaustive requirement extraction: Scans RFCs, Japanese PRDs, OpenAPI/protobuf schemas, user stories, and IaC code using document-type-specific anchors (normative-word grep, full element enumeration, acceptance criteria) and assigns each requirement a sequential ID (S-001...) in a checkbox ledger. - Traceability matrix: Maps each specification clause to its EARS requirement, formal method artifact (TLA+ invariant / Lean property), and test in one shared ledger, making gaps visible as empty cells. - Hard gate enforcement: A PreToolUse hook mechanically blocks downstream formalization steps until every ID is checked off with no numbering gaps. - Use Case: Given an RFC for a protocol you must implement, run this Skill to produce a complete requirements ledger and traceability matrix before writing any TLA+ spec, so no MUST requirement silently disappears from the implementation. ## Quick Start Extract all requirements from this RFC specification into a numbered checklist ledger with a traceability matrix.

Frequently Asked Questions about loopeng-extract

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

FAQPage Schema
How do I extract requirements from an RFC specification?

Grep the RFC for RFC 2119 normative words (MUST, SHALL, SHOULD, MAY and negations) using case-sensitive word-boundary matching, then assign each hit a sequential ID in a checkbox ledger. Convert each entry into one EARS sentence and check it off when done.

How do I ensure requirement extraction is complete before formal verification?

Use a numbered checklist ledger where every specification unit gets an ID, and build a traceability matrix mapping each clause to its EARS requirement, formal method artifact, and test. Missing requirements appear as empty cells, and a hook blocks formalization until all IDs are checked.

Does requirement extraction work for non-RFC documents like PRDs or OpenAPI specs?

Yes, the scanning anchor changes by document type. Japanese PRDs use normative phrases like "shall/must" expressions, OpenAPI and protobuf specs are enumerated fully via parsers, and user stories are classified per acceptance criterion.

Why do formally verified implementations still miss requirements?

TLA+ and Lean verify only the requirements they receive; abstraction deliberately discards details like field validation, creating blind spots at layer boundaries. Without an independent extraction ledger, omitted requirements never appear in any downstream layer.

When should I not use exhaustive requirement extraction?

Avoid narrowing scope yourself when a specification is large; instead split it into MECE parts and process them sequentially. The method is designed for completeness, so informal tasks without a source specification only need the spoken request written down as the starting point.