Getting it into your agent
One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.
npx agentmods add commands/kewton/commandmate/multi-stage-design-reviewgit clone --depth 1 https://github.com/Kewton/CommandMateWrote this? Show the measurements
A badge with what this costs and how it scanned, read live from this page, so it follows the numbers instead of freezing them. Markdown for a README, HTML for a documentation site or a project page.
[](https://agentmods.dev/commands/kewton/commandmate/multi-stage-design-review)<a href="https://agentmods.dev/commands/kewton/commandmate/multi-stage-design-review"><img src="https://agentmods.dev/badge/commands/kewton/commandmate/multi-stage-design-review.svg" alt="Measured on agentmods" height="20"></a>What it costs to keep this loaded
Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.
| Model | Per session | Once invoked |
|---|---|---|
| Fable 5 | $0.00034 | $0.03784 |
| Opus 5 | $0.00017 | $0.01892 |
| Sonnet 5 | $0.00007 | $0.00757 |
| Haiku 4.5 | $0.00003 | $0.00378 |
Grade A, and why
multi-stage-design-review scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 5d ago.
A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.
Nothing flagged
None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.
How it starts
The opening of the file, as written. The whole thing — 374 lines — stays where its author put it; the contents beside it link to each section on GitHub.
マルチステージ設計レビューコマンド
概要
4段階のアーキテクチャレビューとその指摘事項対応を自動で実行するコマンドです。各段階でレビュー→対応のサイクルを回し、設計方針書の品質を段階的に向上させます。
重要: このコマンドは設計方針書のレビューと改善を目的としています。 ソースコードの実装は行いません。レビュー結果は設計方針書に反映されます。
使用方法
/multi-stage-design-review [Issue番号]
/multi-stage-design-review [Issue番号] --skip-stage=3,4
例:
/multi-stage-design-review 76 # 全4段階を実行
/multi-stage-design-review 76 --skip-stage=4 # セキュリティレビューをスキップ
実行内容
あなたはマルチステージレビューの統括者です。4段階のレビューサイクルを順次実行し、各段階で指摘事項を対応してから次の段階に進みます。
パラメータ
- issue_number: 対象Issue番号(必須)
- skip_stage: スキップするステージ番号(カンマ区切り)
サブエージェント/レビュアー指定
| ステージ | レビュアー | 理由 |
|---|---|---|
| Stage 1-2(通常・整合性) | Claude opus(サブエージェント) | 品質判断にOpus必要 |
| Stage 3-4(影響分析・セキュリティ) | Codex(commandmatedev --agent codex 経由) |
異なるモデルによるクロスレビュー |
| 指摘反映(全ステージ) | sonnet(継承) / Codex内で直接反映 | 設計方針書更新 |
⚠ 重要: Stage 3-4 は必ず
commandmatedev send ... --agent codexでCodexに委譲すること。 Claude サブエージェント(Agent tool)で代替実行してはならない。 commandmatedev が利用不可の場合はユーザーに報告して中断すること。
レビューステージ
| Stage | レビュー種別 | フォーカス | 目的 |
|---|---|---|---|
| 1 | 通常レビュー | 設計原則 | SOLID/KISS/YAGNI/DRY準拠確認+発見可能性(discoverability) |
| 2 | 整合性レビュー | 整合性 | 設計書と実装の整合性確認 |
| 3 | 影響分析レビュー | 影響範囲 | 変更の波及効果分析 |
| 4 | セキュリティレビュー | セキュリティ | OWASP Top 10準拠確認 |
実行フェーズ
Phase 0: 初期設定
0-1. TodoWriteで作業計画作成
- [ ] Stage 1: 通常レビュー
- [ ] Stage 1: 指摘事項対応
- [ ] Stage 2: 整合性レビュー
- [ ] Stage 2: 指摘事項対応
- [ ] Stage 3: 影響分析レビュー
- [ ] Stage 3: 指摘事項対応
- [ ] Stage 4: セキュリティレビュー
- [ ] Stage 4: 指摘事項対応
- [ ] 最終確認
0-2. ディレクトリ構造作成
mkdir -p dev-reports/issue/{issue_number}/multi-stage-design-review
Stage 1: 通常レビュー(設計原則)
Stage 1 の必須観点(focus_area: 設計原則):
- SOLID / KISS / YAGNI / DRY 準拠
- 発見可能性(discoverability): この機能・この判定結果を、運用者はどの層で・どのコマンドで知るか?
サーバーログにしか出ない判定・抑止・自動アクションはないか?(判定は理由コードつきで
capture --json/wait/task show等の運用者が読む層に露出すること。 詳細: docs/design/discoverability-principle.md)
What this file has done since we first saw it
Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.
- 5d ago First seen · 374 lines · 34 tokens per session scan A 75aba4b6e36b
multi-stage-design-review is a command published in the GitHub repository Kewton/CommandMate (39 stars, last pushed today), licensed MIT. It adds 34 tokens to every session and 3,784 once invoked, about $0.0002 per session on Opus 5. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-30.
Other commands, from other repositories
crow-image
Show an image in Crow's Images panel — pass the file path as the argument.
release-notes
Draft release notes, recommend and apply a version bump, and update CHANGELOG.md for the next dormouse release by analyzing all merge commits and squash-merged PRs since the last release tag. Used as step 2 of the release checklist in docs/specs/deploy.md.
build-docs
Build and serve the docs site locally with Jekyll.
release
Cut a domux release — tag HEAD, push, watch the GH Action.
review-session-log
Use when the user asks what a session did, wants to review session history, or needs to find errors in a session log.
setup
Install & set up newsline — rotating regional news in your status line (keeps your existing status line).