speckit-adlc-architect-init

Identify architectural uncertainties in brownfield codebases and generate ADR drafts.

Updated Apr 21, 2026
One-click install
npx skills add https://github.com/chaimt/GarminShabbatMode --skill speckit-adlc-architect-init
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: speckit-adlc-architect-init
Source: https://github.com/chaimt/GarminShabbatMode/tree/main/.cursor/skills/speckit-adlc-architect-init
Command: npx skills add https://github.com/chaimt/GarminShabbatMode --skill speckit-adlc-architect-init

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

Brownfield codebases often lack explicit architectural decisions. This Skill analyzes an existing repository to surface inferred ADRs and guidance for clarification.

Core Features & Use Cases

  • Discovers architectural decisions from code structure to draft ADRs.
  • Outputs ADRs to {REPO_ROOT}/.specify/drafts/adr.md for review.
  • Provides a validation path via /architect.clarify to confirm findings.
  • Supports sub-system decomposition and documentation deduplication when present.

Quick Start

Run the architect init workflow on a brownfield codebase to generate ADRs and prepare them for clarification.

Frequently Asked Questions about speckit-adlc-architect-init

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

FAQPage Schema
How do I generate ADRs from an existing brownfield codebase?

To generate ADRs from a brownfield codebase, apply the architect init workflow to analyze the repository structure and surface inferred architectural decisions. The discovered decisions are drafted as ADRs to {REPO_ROOT}/.specify/drafts/adr.md for review.

How do I prepare architectural clarifications for validation in a monolithic repository?

Preparing architectural clarifications involves identifying uncertainties in a multi-module repository and drafting inferred ADRs. These drafts are then validated through the /architect.clarify path to confirm the discovered architectural findings.

Can I reverse-engineer architectural decisions from multi-module code without documentation?

Yes, you can reverse-engineer architectural decisions from undocumented multi-module code by analyzing the codebase structure. This process discovers implicit decisions, drafts them into ADRs, and supports sub-system decomposition to deduplicate existing documentation when present.

What is the best way to document inferred architectural decisions for validation?

The best way to document inferred architectural decisions is generating ADR drafts from the codebase analysis and writing them to {REPO_ROOT}/.specify/drafts/adr.md. This provides a structured validation path via /architect.clarify to confirm findings.

Does the ADR generation process support sub-system decomposition in monolithic repositories?

Yes, the ADR generation process supports sub-system decomposition and documentation deduplication when present in monolithic repositories. It analyzes the code structure to surface inferred architectural decisions and outputs them for clarification.