my-take-dev GitHub avatar

inspired-mino-design-skills

my-take-dev

An unofficial, experimental suite of AI Skills that reconstructs design principles from Mino's public materials into repeatable, evidence-based decision workflows.

Stars

270

7-day growth

No data

Forks

9

Open issues

1

License

No data

Last updated

2026-07-15

Why it is worth attention

It uniquely formalizes design heuristics into composable, platform-aware AI Skills while explicitly separating human judgment from automated decisions, with rigorous structural validation across Windows, Linux, and macOS.

Who it is for

  • Software developers and architects applying structured design methods
  • AI/LLM agent developers seeking repeatable design guidance
  • Engineering teams adopting contract-driven and domain-driven design
  • Technical leads evaluating design decisions with verifiable evidence

Use cases

  • Clarifying ambiguous requirements by framing the problem before implementation
  • Auditing a domain model for missing concepts, states, or constraints
  • Translating natural-language requirements into testable pre/post conditions and contracts
  • Designing architecture quality strategies with explicit trade-offs and recovery plans

Strengths

  • Grounded in published design principles, not fictional best practices
  • Cross-platform structural validation (Linux, Windows, macOS) with versioned test fixtures
  • Clear separation of factual evidence, interpretation, and human-owned decisions
  • Composable Skills that apply only the minimum needed scope for each task

Considerations

  • Current experimental state (v0.9.0) has not passed behavioral release; only structural validation is partially complete
  • Native macOS fixture runner fails, blocking full platform parity
  • Requires specific environment setup (Codex CLI with .agents/skills) and is not yet ready for production use

README quick start

mino-drive-inspired-design-skills

本リポジトリは、ミノ駆動氏の公開資料を基に、設計判断の原則をAI Skillとして再構成する非公式プロジェクトです。ミノ駆動氏および所属組織による監修・承認・推奨を受けたものではありません。

目指すのは人物の口調や結論の模倣ではありません。問題、目的、文脈、要件、契約、モデル、品質、公開境界を確認し、根拠のある判断と検証可能な成果物を繰り返し作れるようにすることです。公開資料から直接抽出した原則、反復可能なSkillにするためsuiteが追加したschema・gate・workflow、versioningやvalidator等のrepository policyは区別して記載します。

現在の状態

Status: Experimental / Preview
Suite version: 0.9.0
Structural validation: fail on the native macOS fixture runner; pass on executed Linux and PowerShell-over-WSL layers
macOS structural support: implemented; native macOS /bin/bash executed / fail (run 29397674053, exit 2)
Targeted behavioral evidence: not executed
Behavioral release: not ready

ここでいうbehavioral release(行動再現性を確認した安定版判定)は、権限を持つevaluation ownerがfrozenにしたversioned caseと隔離oracleをfresh contextで繰り返し実行し、代表case・negative case・回帰・required platformのRelease gateを満たした状態です。Experimental / Previewとして配布可能であることと、stable releaseとして承認済みであることを分けます。権限を持つmaintainerが全Evidenceを確認するまでstable releaseとは扱いません。

プロダクト価値(利用者・事業が得る成果)と品質portfolio(優先する品質、制約、意図的に最適化しない品質の組合せ)の最終判断も、AIではなく権限を持つ人間が所有します。

native macOS Evidenceは、head e47aaafb74a27cf2cc7d4bc9c64f74d1933f10dbのworkflow run 29397674053、job Native macOS /bin/bash(job ID 87294760529)で取得済みです。環境はmacOS 15.7.7、image macos-15-arm64 version 20260706.0213.1RUNNER_ARCH=ARM64でした。validate-suite.shはpassしましたが、fixture runnerはsolver-nested-metadataのportable rewriteで失敗し、jobはexit 2でした。

macOSはBash構造validator、fixture runner、text-format検査を対応範囲に含めます。共用scriptは標準/bin/bash 3.2で実行できるsubsetへ制限しています。上記failureを修正したheadでnative macOS jobがgreenになるまで、platform parityやreleaseをpassにしません。

クイックスタート

Skill routingの適用

このsuiteは技術非依存の設計原則を扱います。プログラミング言語、framework、tool固有のSkillを併用する場合は、主成果物に最も合うmino Skillを基本workflowとし、必要な技術差分だけを追加します。現在のrouting規則は、maintainerが責任を持って採用・保守しています。

Skill compositionを利用側のAGENTS.mdへ記載する例:

# Skill Composition

When a request matches multiple Skills:

- Use the Skill that best matches the primary outcome as the basic workflow.
- Add only relevant language-, framework-, or tool-specific Skills to supplement that workflow.
- Let the basic Skill control scope, changes, validation, and the final response; specialized Skills provide their domain-specific guidance.
- Preserve every applicable Skill's exclusions, hard gates, and safety constraints.
- Follow the user's explicitly named Skills and do

Description

ミノ駆動氏のWebに掲載されている資料から作成した開発スキル

Related repositories

Similar projects matched by category, topics, and programming language.

jamesob
Featured
jamesob GitHub avatar

local-llm

A comprehensive guide for building and configuring a high-end local machine to run state-of-the-art LLMs, with detailed hardware choices, BIOS tuning, and Docker-based model serving.

Shell
1,520
iptvorganization
iptvorganization GitHub avatar

TV-AutoBuild

An automated GitHub Actions workflow that continuously monitors multiple source repositories, builds consistently signed release APKs with deterministic versioning, and supports scheduled, manual, and event-driven triggers.

Shell
165
aws-samples
aws-samples GitHub avatar

sample-specship

Spec-driven autonomous engineering workflow for AI coding agents: recon → plan → build → validate → ship — with TDD, adversarial validation, and anti-slop quality gates. Packaged as a Kiro Power.

Shell
160