Agent-Skills.md

Agent Skills: ASR Extractor

Internal Moonshot Architecture stage for extracting architecturally significant requirements and quality attribute scenarios from normalized requirements.

UncategorizedID: munlucky/claude-settings/asr-extractor

Repository

munlucky/claude-settings
munlucky

Install this agent skill to your local

pnpm dlx add-skill https://github.com/munlucky/moonshot-relay/tree/HEAD/skills/asr-extractor

Skill Files

Browse the full folder contents for asr-extractor.

Download Skill

Loading file tree…

skills/asr-extractor/SKILL.md

Skill Metadata

Name
asr-extractor
Description
Internal Moonshot Architecture stage for extracting architecturally significant requirements and quality attribute scenarios from normalized requirements.

ASR Extractor

Role

Extract architecturally significant requirements from REQUIREMENT_INVENTORY.md, PRD evidence, and Brownfield constraints.

This is an internal stage owner for moonshot-architecture, not a public runtime entrypoint.

Inputs

  • REQUIREMENT_INVENTORY.md
  • PRD, SPEC, or Brownfield evidence summary
  • Known constraints, risks, and quality attributes

Flow

  1. Identify requirements that materially affect structure, runtime behavior, data boundaries, security, performance, reliability, operability, or integration contracts.
  2. Assign stable ASR-001 style IDs and link each ASR to one or more REQ-001 style IDs.
  3. Write quality attribute scenarios with stimulus, environment, response, and response measure.
  4. Record rejected ASR candidates when a requirement is product-important but not architecture-significant.
  5. Preserve traceability links for downstream option generation and ADR/C4 writing.

Hard Stops

  • Do not label every requirement as an ASR.
  • Do not create ASRs without requirement IDs.
  • Do not invent quality measures when the source is silent; mark the gap explicitly.
  • Do not expose raw MemoryGraph, KG, ontology, log, transcript, browser, or secret-like data.

Required Evidence

  • Source requirement IDs.
  • ASR IDs and rationale.
  • Quality attribute scenarios.
  • Verification signal or explicit verification gap for each ASR.