superpowers-mcp 6.3.1 → 6.3.3
This diff represents the content of publicly available package versions that have been released to one of the supported registries. The information contained in this diff is provided for informational purposes only and reflects changes between package versions as they appear in their respective public registries.
- package/README.ja.md +33 -8
- package/README.ko.md +33 -8
- package/README.md +33 -8
- package/README.zh-TW.md +32 -7
- package/out/server.js +39 -15
- package/package.json +1 -1
- package/skills/subagent-driven-development/SKILL.md +56 -2
- package/skills/subagent-driven-development/implementer-prompt.md +5 -2
- package/skills/using-superpowers/SKILL.md +2 -0
- package/skills/using-superpowers/references/devin-tools.md +41 -0
- package/skills/using-superpowers/references/opencode-tools.md +32 -0
- package/skills/writing-plans/SKILL.md +24 -0
- package/skills/writing-plans/skeleton-first-plans.md +131 -0
- package/skills/writing-skills/render-graphs.js +7 -6
package/README.ja.md
CHANGED
|
@@ -2,28 +2,38 @@
|
|
|
2
2
|
|
|
3
3
|
[English](README.md) | [繁體中文](README.zh-TW.md) | [日本語](README.ja.md) | [한국어](README.ko.md)
|
|
4
4
|
|
|
5
|
-
[](https://github.com/Poseidoncode/superpowers-mcp)
|
|
6
6
|
[](LICENSE)
|
|
7
7
|
|
|
8
|
-
|
|
8
|
+
このドキュメントは、Superpowers スキルライブラリと自律型ワークフローを、独立した高パフォーマンスかつ安全な **Model Context Protocol (MCP)** サーバーにパッケージ化した使用説明書です。
|
|
9
9
|
|
|
10
10
|
---
|
|
11
11
|
|
|
12
12
|
## 🚀 インストールと使用方法
|
|
13
13
|
|
|
14
|
-
###
|
|
14
|
+
### サポート環境とエージェントプラットフォーム
|
|
15
15
|
|
|
16
|
-
**Antigravity**、**Cursor**、**VSCode
|
|
16
|
+
- **MCP ネイティブ AI エディター & IDE**: **Antigravity (AGY)**、**Cursor**、**VSCode**、**Windsurf**、**Claude Desktop / Claude Code**。
|
|
17
|
+
- **CLI ツール & 自律型エージェント**: **Devin CLI**、**Hermes Agent**、**OpenCode**、**Kimi CLI**、**Pi CLI**、**Gemini CLI**。
|
|
18
|
+
|
|
19
|
+
### 提供される MCP 機能
|
|
20
|
+
|
|
21
|
+
| プロトコル機能 | 項目 | 説明 |
|
|
22
|
+
| :--- | :--- | :--- |
|
|
23
|
+
| **Tools** | `list_skills`, `read_skill` | 14 種類の Superpowers スキルをオンデマンドで検索・読み込み。 |
|
|
24
|
+
| **Prompts** | `session-start`, `sdd-implementer`, `sdd-task-reviewer`, `sdd-re-review`, `spec-reviewer`, `plan-reviewer` | IDE のプロンプト一覧から即座に使用可能なレビュー・サブエージェント協調プロンプト。 |
|
|
25
|
+
| **Resources** | `skill://superpowers/<skill-name>` | MCP 規格に準拠した URI ベースのスキルアクセス。 |
|
|
17
26
|
|
|
18
27
|
### AI エージェントとの対話
|
|
19
28
|
|
|
20
|
-
インストールまたは設定が完了すると、AI
|
|
29
|
+
インストールまたは設定が完了すると、AI エージェントが自動的に `Superpowers Skills` および `Prompts` を認識して呼び出せるようになります。
|
|
21
30
|
|
|
22
|
-
|
|
31
|
+
**質問例:**
|
|
23
32
|
|
|
24
33
|
- 「すべての superpowers スキルを一覧表示して」
|
|
25
34
|
- 「read_skill で brainstorming スキルを読み込んで、この機能の実装を分析して」
|
|
26
|
-
- 「session-start
|
|
35
|
+
- 「session-start プロンプトを適用して」(Superpowers のコンテキストをセッションに注入)
|
|
36
|
+
- 「subagent-driven-development で docs/plans/feature-plan.md の計画を実行して」
|
|
27
37
|
|
|
28
38
|
---
|
|
29
39
|
|
|
@@ -128,7 +138,22 @@
|
|
|
128
138
|
|
|
129
139
|
## 🆕 最近の更新
|
|
130
140
|
|
|
131
|
-
### v6.3.
|
|
141
|
+
### v6.3.2(最新)
|
|
142
|
+
|
|
143
|
+
- **writing-plans — 2 つのプラン形状(Two Plan Shapes)と骨格優先(Skeleton-First)**:
|
|
144
|
+
- `skills/writing-plans/SKILL.md` に **Two Plan Shapes** ルーターを追加(`task-by-task` デフォルト vs `skeleton-first` 代替案)。
|
|
145
|
+
- 新規 [`skills/writing-plans/skeleton-first-plans.md`](skills/writing-plans/skeleton-first-plans.md) で Walking Skeleton(Task 1 で全サブシステムを貫通する最小稼働スライスを作成)、契約型タスク(Task Contracts、コードを直接書かずに厳密な Consumes/Produces インターフェースと観察可能な検証基準を定義)、`Tier: mechanical | judgment` タグを定義。
|
|
146
|
+
- **subagent-driven-development (SDD) — ウェーブディスパッチ(Wave Dispatch)と並列 Worktree プロトコル**:
|
|
147
|
+
- skeleton-first プランに対して **DISPATCH PLAN** を生成し、ファイルの競合がないタスクをウェーブとして並列ディスパッチ。
|
|
148
|
+
- **並列 Worktree プロトコル(Parallel Worktree Protocol)**:独立した `.worktrees/task-<N>` で並行タスクを実行し、プラン順序で順次統合。競合時は自動 rebase と implementer 再開で修正。
|
|
149
|
+
- Step 5 に完了後の `Plan holds` / `Amendment:` チェック行を追加し、進行中タスクの整合性を保ちながら後続タスクに更新された契約を伝播。
|
|
150
|
+
- **SDD — Tier 駆動モデルディスパッチ**:
|
|
151
|
+
- SDD ディスパッチャーと `implementer-prompt.md` が `Tier:` 指定(`mechanical` → 経済的な軽量モデル、`judgment` → 標準モデル)を直接適用し、トークンを節約。
|
|
152
|
+
- **writing-skills — バイナリ実行セキュリティ強化(`render-graphs.js`)**:
|
|
153
|
+
- `execSync` から `execFileSync('dot', ['-Tsvg'], ...)` に移行してシェルインジェクションを根絶。10MB バッファ上限、Windows CRLF 対応(`\r?\n`)、`winget` インストール案内を追加。
|
|
154
|
+
- **テストと検証**:MCP プロトコル、セキュリティ、SDD Bash(11 アサーション)、PowerShell(70 アサーション)、Graphviz レンダリングテストのすべてが 100% 合格。
|
|
155
|
+
|
|
156
|
+
### v6.3.1
|
|
132
157
|
|
|
133
158
|
- **SDD ワークスペース所有権マーカーと物理パス正規化**:`sdd-workspace`(Bash)と `sdd-workspace.ps1`(PowerShell)は `plan-path` マーカーと物理パス正規化(`pwd -P` および動的 `pwd` 検出)を使用。同名のプラン(`docs/alpha/plan.md` と `docs/beta/plan.md` など)が衝突せず個別の `.superpowers/sdd/` ワークスペースに分離されます。
|
|
134
159
|
- **SDD レビューパッケージ範囲ガード**:`review-package` および `review-package.ps1` は `git merge-base --is-ancestor BASE HEAD` とコミット数を検証し、空や逆転した範囲による誤判定(false-pass)を防止。
|
package/README.ko.md
CHANGED
|
@@ -2,28 +2,38 @@
|
|
|
2
2
|
|
|
3
3
|
[English](README.md) | [繁體中文](README.zh-TW.md) | [日本語](README.ja.md) | [한국어](README.ko.md)
|
|
4
4
|
|
|
5
|
-
[](https://github.com/Poseidoncode/superpowers-mcp)
|
|
6
6
|
[](LICENSE)
|
|
7
7
|
|
|
8
|
-
이 문서는
|
|
8
|
+
이 문서는 Superpowers 스킬 라이브러리와 자율 에이전트 워크플로우를 독립적이고 고성능이며 안전한 **Model Context Protocol (MCP)** 서버로 패키징한 사용 지침을 요약한 것입니다.
|
|
9
9
|
|
|
10
10
|
---
|
|
11
11
|
|
|
12
12
|
## 🚀 설치 및 사용 방법
|
|
13
13
|
|
|
14
|
-
### 지원 환경
|
|
14
|
+
### 지원 환경 및 에이전트 플랫폼
|
|
15
15
|
|
|
16
|
-
**Antigravity**, **Cursor**, **VSCode**
|
|
16
|
+
- **MCP 네이티브 AI 편집기 & IDE**: **Antigravity (AGY)**, **Cursor**, **VSCode**, **Windsurf**, **Claude Desktop / Claude Code**.
|
|
17
|
+
- **CLI 도구 & 자율 에이전트**: **Devin CLI**, **Hermes Agent**, **OpenCode**, **Kimi CLI**, **Pi CLI**, **Gemini CLI**.
|
|
18
|
+
|
|
19
|
+
### 제공되는 MCP 프로토콜 기능
|
|
20
|
+
|
|
21
|
+
| 프로토콜 기능 | 포함 항목 | 설명 |
|
|
22
|
+
| :--- | :--- | :--- |
|
|
23
|
+
| **Tools** | `list_skills`, `read_skill` | 14개의 Superpowers 스킬을 온디맨드로 검색 및 로드합니다. |
|
|
24
|
+
| **Prompts** | `session-start`, `sdd-implementer`, `sdd-task-reviewer`, `sdd-re-review`, `spec-reviewer`, `plan-reviewer` | IDE 프롬프트 목록에서 즉시 사용할 수 있는 대립 검토 및 서브에이전트 조율 프롬프트. |
|
|
25
|
+
| **Resources** | `skill://superpowers/<skill-name>` | MCP 규격에 맞는 URI 기반 스킬 직접 접근. |
|
|
17
26
|
|
|
18
27
|
### AI 에이전트와 대화하기
|
|
19
28
|
|
|
20
|
-
설치 또는 구성이 완료되면 AI
|
|
29
|
+
설치 또는 구성이 완료되면 AI 에이전트가 `Superpowers Skills` 및 `Prompts`를 자동으로 인식하고 호출할 수 있습니다.
|
|
21
30
|
|
|
22
|
-
|
|
31
|
+
**질문 예시:**
|
|
23
32
|
|
|
24
33
|
- "모든 superpowers 스킬을 나열해줘"
|
|
25
34
|
- "read_skill을 사용하여 brainstorming 스킬을 읽고, 이 기능의 구현을 분석해줘"
|
|
26
|
-
- "session-start 프롬프트를 적용해줘" (
|
|
35
|
+
- "session-start 프롬프트를 적용해줘" (Superpowers 전체 워크플로우 컨텍스트 초기화)
|
|
36
|
+
- "subagent-driven-development로 docs/plans/feature-plan.md 계획을 실행해줘"
|
|
27
37
|
|
|
28
38
|
---
|
|
29
39
|
|
|
@@ -128,7 +138,22 @@
|
|
|
128
138
|
|
|
129
139
|
## 🆕 최근 업데이트
|
|
130
140
|
|
|
131
|
-
### v6.3.
|
|
141
|
+
### v6.3.2 (최신)
|
|
142
|
+
|
|
143
|
+
- **writing-plans — 2가지 계획 형태(Two Plan Shapes)와 스켈레톤 우선(Skeleton-First)**:
|
|
144
|
+
- `skills/writing-plans/SKILL.md`에 **Two Plan Shapes** 라우터를 추가(`task-by-task` 기본값 vs `skeleton-first` 대안).
|
|
145
|
+
- 새 [`skills/writing-plans/skeleton-first-plans.md`](skills/writing-plans/skeleton-first-plans.md)에서 Walking Skeleton(Task 1에서 전체 하위 시스템을 얇게 관통하는 실행 슬라이스 구축), 계약 기반 작업(Task Contracts, 코드 스크립트 대신 엄격한 Consumes/Produces 인터페이스 및 관찰 가능한 성공 기준 명시), `Tier: mechanical | judgment` 태그 정의.
|
|
146
|
+
- **subagent-driven-development (SDD) — 웨이브 디스패치(Wave Dispatch) 및 병렬 Worktree 프로토콜**:
|
|
147
|
+
- skeleton-first 계획에 대해 **DISPATCH PLAN**을 생성하여 파일 충돌이 없는 작업을 웨이브 단위로 병렬 디스패치.
|
|
148
|
+
- **병렬 Worktree 프로토콜(Parallel Worktree Protocol)**: 독립된 `.worktrees/task-<N>`에서 동시 작업을 실행하고 계획 순서대로 순차 병합. 충돌 시 자동 rebase 후 implementer를 재개하여 자체 해결.
|
|
149
|
+
- Step 5에 완료 후 `Plan holds` / `Amendment:` 점검 라인을 추가하여 진행 중인 작업의 무결성을 유지하면서 후속 작업에 변경된 계약 전파.
|
|
150
|
+
- **SDD — Tier 기반 모델 디스패치**:
|
|
151
|
+
- SDD 디스패처 및 `implementer-prompt.md`가 `Tier:` 지정(`mechanical` → 경제적인 경량 모델, `judgment` → 표준 모델)을 즉시 적용하여 토큰 낭비 방지.
|
|
152
|
+
- **writing-skills — 바이너리 실행 보안 강화(`render-graphs.js`)**:
|
|
153
|
+
- `execSync` 대신 `execFileSync('dot', ['-Tsvg'], ...)`로 전환하여 쉘 인젝션 위험 근절. 10MB 버퍼 제한, Windows CRLF(`\r?\n`) 호환 및 `winget` 설치 안내 추가.
|
|
154
|
+
- **테스트 및 검증**: MCP 프로토콜, 보안, SDD Bash(11개 어서션), PowerShell(70개 어서션), Graphviz 렌더링 테스트 100% 통과.
|
|
155
|
+
|
|
156
|
+
### v6.3.1
|
|
132
157
|
|
|
133
158
|
- **SDD 워크스페이스 소유권 마커 및 물리 경로 정규화**: `sdd-workspace`(Bash) 및 `sdd-workspace.ps1`(PowerShell)은 `plan-path` 마커와 물리 경로 정규화(`pwd -P` 및 동적 `pwd` 감지)를 사용. 동일한 이름의 계획 파일(`docs/alpha/plan.md` 및 `docs/beta/plan.md` 등)이 충돌 없이 별도의 `.superpowers/sdd/` 워크스페이스로 격리됩니다.
|
|
134
159
|
- **SDD 리뷰 패키지 범위 가드**: `review-package` 및 `review-package.ps1`은 `git merge-base --is-ancestor BASE HEAD`와 커밋 수를 검증하여 비어 있거나 역전된 범위로 인한 오탐(false-pass)을 방지합니다.
|
package/README.md
CHANGED
|
@@ -2,28 +2,38 @@
|
|
|
2
2
|
|
|
3
3
|
[English](README.md) | [繁體中文](README.zh-TW.md) | [日本語](README.ja.md) | [한국어](README.ko.md)
|
|
4
4
|
|
|
5
|
-
[](https://github.com/Poseidoncode/superpowers-mcp)
|
|
6
6
|
[](LICENSE)
|
|
7
7
|
|
|
8
|
-
This document summarizes the information and usage instructions for packaging the
|
|
8
|
+
This document summarizes the information and usage instructions for packaging the Superpowers skills and autonomous workflow system into an independent, high-performance, and secure **Model Context Protocol (MCP)** server.
|
|
9
9
|
|
|
10
10
|
---
|
|
11
11
|
|
|
12
12
|
## 🚀 How to Install and Use
|
|
13
13
|
|
|
14
|
-
### Supported Environments
|
|
14
|
+
### Supported Environments & Harnesses
|
|
15
15
|
|
|
16
|
-
**Antigravity**, **Cursor**, **VSCode**,
|
|
16
|
+
- **MCP-Native AI Editors & IDEs**: **Antigravity (AGY)**, **Cursor**, **VSCode**, **Windsurf**, **Claude Desktop / Claude Code**.
|
|
17
|
+
- **CLI & Autonomous Agents**: **Devin CLI**, **Hermes Agent**, **OpenCode**, **Kimi CLI**, **Pi CLI**, **Gemini CLI**.
|
|
18
|
+
|
|
19
|
+
### MCP Capabilities Provided
|
|
20
|
+
|
|
21
|
+
| Protocol Feature | Items / Capabilities | Description |
|
|
22
|
+
| :--- | :--- | :--- |
|
|
23
|
+
| **Tools** | `list_skills`, `read_skill` | Discover and load any of the 14 Superpowers skills on demand. |
|
|
24
|
+
| **Prompts** | `session-start`, `sdd-implementer`, `sdd-task-reviewer`, `sdd-re-review`, `spec-reviewer`, `plan-reviewer` | Ready-to-use adversarial review and subagent orchestrator prompts in your IDE prompt list. |
|
|
25
|
+
| **Resources** | `skill://superpowers/<skill-name>` | Direct URI-based skill access for MCP-compliant hosts. |
|
|
17
26
|
|
|
18
27
|
### Chatting with the AI Agent
|
|
19
28
|
|
|
20
|
-
Once installed or configured, your AI Agent
|
|
29
|
+
Once installed or configured, your AI Agent will automatically discover and invoke `Superpowers Skills` and `Prompts`.
|
|
21
30
|
|
|
22
|
-
**
|
|
31
|
+
**Example prompts to your agent:**
|
|
23
32
|
|
|
24
33
|
- "List all superpowers skills"
|
|
25
34
|
- "Use read_skill to read the brainstorming skill, and then help me analyze the implementation of this feature"
|
|
26
|
-
- "Apply the session-start prompt" (
|
|
35
|
+
- "Apply the session-start prompt" (Initializes the full Superpowers workflow context)
|
|
36
|
+
- "Run Subagent-Driven Development for docs/plans/feature-plan.md"
|
|
27
37
|
|
|
28
38
|
---
|
|
29
39
|
|
|
@@ -128,7 +138,22 @@ These skills are designed for orchestrating complex meta-execution patterns with
|
|
|
128
138
|
|
|
129
139
|
## 🆕 Recent Updates
|
|
130
140
|
|
|
131
|
-
### v6.3.
|
|
141
|
+
### v6.3.2 (Latest)
|
|
142
|
+
|
|
143
|
+
- **writing-plans — Two Plan Shapes Router & Skeleton-First Plans**:
|
|
144
|
+
- `skills/writing-plans/SKILL.md` introduces the **Two Plan Shapes** router (`task-by-task` default vs `skeleton-first` alternative) to determine architecture upfront.
|
|
145
|
+
- New [`skills/writing-plans/skeleton-first-plans.md`](skills/writing-plans/skeleton-first-plans.md) defines the Walking Skeleton pattern (Task 1 builds the thinnest running end-to-end slice across all subsystems), Task Contracts (strict interfaces and observable behaviors without code scripts), and deliberate `Tier: mechanical | judgment` tagging.
|
|
146
|
+
- **subagent-driven-development (SDD) — Wave Dispatch & Parallel Worktree Protocol**:
|
|
147
|
+
- SDD controller performs **Dispatch Plan** scanning on skeleton-first plans, grouping mutually file-disjoint tasks into waves for concurrent dispatch.
|
|
148
|
+
- **Parallel Worktree Protocol**: Runs concurrent tasks in dedicated `.worktrees/task-<N>` worktrees, with sequential plan-order merges and automatic rebase-to-fix loops on merge conflict or test regression.
|
|
149
|
+
- Step 5 adds the post-completion `Plan holds` / `Amendment:` check line, ensuring in-flight tasks complete cleanly while downstream tasks inherit updated plan contracts.
|
|
150
|
+
- **SDD — Tier-Driven Model Selection**:
|
|
151
|
+
- SDD dispatcher and `implementer-prompt.md` respect task `Tier:` markings (`mechanical` → fastest/cheapest tier; `judgment` → standard tier), saving tokens without redundant re-adjudication.
|
|
152
|
+
- **writing-skills — Binary Execution Hardening (`render-graphs.js`)**:
|
|
153
|
+
- Replaced `execSync` shell execution with direct `execFileSync('dot', ['-Tsvg'], ...)` to eliminate shell injection risks. Added 10MB buffer limits, Windows CRLF support (`\r?\n`), and Windows `winget` installation guidance.
|
|
154
|
+
- **Tests & Verification**: Full regression suites pass 100% across MCP protocol, Security Edge Cases, SDD Bash (11 assertions), PowerShell (70 assertions), and Graphviz rendering.
|
|
155
|
+
|
|
156
|
+
### v6.3.1
|
|
132
157
|
|
|
133
158
|
- **SDD Ownership Markers & Path Normalization**: `sdd-workspace` (Bash) and `sdd-workspace.ps1` (PowerShell) now manage plan-scoped workspaces using `plan-path` markers with canonical physical path normalization (`pwd -P` and dynamic `pwd` detection). Same-basename plans (e.g. `docs/alpha/plan.md` vs `docs/beta/plan.md`) safely disambiguate into distinct `.superpowers/sdd/` workspaces, preventing artifact and ledger overwrites.
|
|
134
159
|
- **SDD Review Package Range Mechanical Guards**: `review-package` and `review-package.ps1` now enforce `git merge-base --is-ancestor BASE HEAD` and `git rev-list --count BASE..HEAD > 0` (exiting with code 3 on error) to reject invalid or empty commit ranges and prevent false-pass review approvals.
|
package/README.zh-TW.md
CHANGED
|
@@ -2,28 +2,38 @@
|
|
|
2
2
|
|
|
3
3
|
[English](README.md) | [繁體中文](README.zh-TW.md) | [日本語](README.ja.md) | [한국어](README.ko.md)
|
|
4
4
|
|
|
5
|
-
[](https://github.com/Poseidoncode/superpowers-mcp)
|
|
6
6
|
[](LICENSE)
|
|
7
7
|
|
|
8
|
-
|
|
8
|
+
本文檔總結了將 Superpowers 技能庫與自主 Agent 工作流架構打包成獨立、高效能且安全加固的 **Model Context Protocol (MCP)** 伺服器之相關資訊與使用說明。
|
|
9
9
|
|
|
10
10
|
---
|
|
11
11
|
|
|
12
12
|
## 🚀 安裝與使用方式
|
|
13
13
|
|
|
14
|
-
###
|
|
14
|
+
### 支援的環境與 Agent 平台
|
|
15
15
|
|
|
16
|
-
**Antigravity**、**Cursor**、**VSCode
|
|
16
|
+
- **MCP 原生 AI 編輯器與 IDE**:**Antigravity (AGY)**、**Cursor**、**VSCode**、**Windsurf**、**Claude Desktop / Claude Code**。
|
|
17
|
+
- **CLI 工具與自主代理環境**:**Devin CLI**、**Hermes Agent**、**OpenCode**、**Kimi CLI**、**Pi CLI**、**Gemini CLI**。
|
|
18
|
+
|
|
19
|
+
### 提供之 MCP 協議功能
|
|
20
|
+
|
|
21
|
+
| 協議功能 | 包含項目 | 說明 |
|
|
22
|
+
| :--- | :--- | :--- |
|
|
23
|
+
| **Tools (工具)** | `list_skills`, `read_skill` | 依需求隨時探索並載入 14 項 Superpowers 技能完整內容。 |
|
|
24
|
+
| **Prompts (提示詞)** | `session-start`, `sdd-implementer`, `sdd-task-reviewer`, `sdd-re-review`, `spec-reviewer`, `plan-reviewer` | 可在 IDE Prompts 列表直接選用之對抗式審查與子代理協調 Prompt 模板。 |
|
|
25
|
+
| **Resources (資源)** | `skill://superpowers/<skill-name>` | 提供符合 MCP 標準的 Direct URI 技能讀取協議。 |
|
|
17
26
|
|
|
18
27
|
### 與 AI Agent 對話
|
|
19
28
|
|
|
20
|
-
安裝或配置完成後,您的 AI Agent
|
|
29
|
+
安裝或配置完成後,您的 AI Agent 將能夠自動識別並調用 `Superpowers Skills` 與 `Prompts`。
|
|
21
30
|
|
|
22
31
|
**您可以這樣提問:**
|
|
23
32
|
|
|
24
33
|
- "列出所有 superpowers 技能"
|
|
25
34
|
- "使用 read_skill 讀取 brainstorming 技能,然後幫我分析這個功能的實作"
|
|
26
|
-
- "套用 session-start prompt"
|
|
35
|
+
- "套用 session-start prompt"(初始化完整的 Superpowers 工作流上下文)
|
|
36
|
+
- "使用 subagent-driven-development 執行 docs/plans/feature-plan.md 的開發計畫"
|
|
27
37
|
|
|
28
38
|
---
|
|
29
39
|
|
|
@@ -128,7 +138,22 @@
|
|
|
128
138
|
|
|
129
139
|
## 🆕 最近更新
|
|
130
140
|
|
|
131
|
-
### v6.3.
|
|
141
|
+
### v6.3.2 (最新版)
|
|
142
|
+
|
|
143
|
+
- **writing-plans — 雙計畫形態(Two Plan Shapes)與骨架優先(Skeleton-First)**:
|
|
144
|
+
- `skills/writing-plans/SKILL.md` 加入 **Two Plan Shapes** 路由機制(`task-by-task` 預設模式 vs `skeleton-first` 骨架優先),在撰寫計畫前提早決定架構模型。
|
|
145
|
+
- 新增 [`skills/writing-plans/skeleton-first-plans.md`](skills/writing-plans/skeleton-first-plans.md) 定義 Walking Skeleton(Task 1 先完成貫穿所有子系統的真實薄切片)、契約式任務(Task Contracts,嚴格定義 Consumes/Produces 接口與可觀測驗收標準,不預寫代碼腳本)與 `Tier: mechanical | judgment` 標記。
|
|
146
|
+
- **subagent-driven-development (SDD) — 波次派發(Wave Dispatch)與並行工作樹協議**:
|
|
147
|
+
- 控制器在衝突掃描階段針對 skeleton-first 計畫生成 **DISPATCH PLAN**,將檔案互斥且無未完成介面依賴的任務組成波次並行派發。
|
|
148
|
+
- **並行工作樹協議(Parallel Worktree Protocol)**:並行任務各於專屬 `.worktrees/task-<N>` 執行,按計畫順序依次合併;若遇衝突或驗證失敗,自動 rebase 並喚醒 implementer 自行修復。
|
|
149
|
+
- Step 5 增加完工後的 `Plan holds` / `Amendment:` 檢查行,確保並行在途任務順利收斂,並將介面變動直接傳遞給後續任務。
|
|
150
|
+
- **SDD — Tier 驅動模型分派**:
|
|
151
|
+
- SDD 控制器與 `implementer-prompt.md` 嚴格遵循計畫中的 `Tier:` 標記(`mechanical` → 快速經濟型模型;`judgment` → 標準中階模型),大幅節省 token 且不再重複審議。
|
|
152
|
+
- **writing-skills — 二進位安全加固(`render-graphs.js`)**:
|
|
153
|
+
- 改用 `execFileSync('dot', ['-Tsvg'], ...)` 徹底杜絕 shell 注入,設置 10MB 緩衝上限,支援 Windows CRLF(`\r?\n`)換行匹配與 `winget` 安裝提示。
|
|
154
|
+
- **測試與驗證**:MCP 協定、邊界安全、SDD Bash(11 項斷言)、PowerShell(70 項斷言)及 Graphviz 渲染測試全數 100% 通過。
|
|
155
|
+
|
|
156
|
+
### v6.3.1
|
|
132
157
|
|
|
133
158
|
- **SDD 工作區歸屬標記與實體路徑正規化**:`sdd-workspace`(Bash)與 `sdd-workspace.ps1`(PowerShell)改用 `plan-path` 歸屬標記與標準化實體路徑(`pwd -P` 與動態 `pwd` 偵測)。同名 plan(如 `docs/alpha/plan.md` 與 `docs/beta/plan.md`)會自動解析為獨立的 `.superpowers/sdd/` 工作區,杜絕產物與進度覆蓋。
|
|
134
159
|
- **SDD 審查包範圍機械守衛**:`review-package` 與 `review-package.ps1` 強制校驗 `git merge-base --is-ancestor BASE HEAD` 與 commit 數量大於 0(錯誤時 exit 3),防止空範圍或倒置範圍造成審查假綠燈。
|