writing-plans
Compare original and translation side by side
🇺🇸
Original
English🇨🇳
Translation
Chinese寫作計劃
Writing Plan
概述
Overview
假設工程師對我們的程式碼庫的背景為零且品味有問題,則編寫全面的實施計劃。記錄他們需要知道的一切:每個任務要接觸哪些文件、程式碼、測試、他們可能需要檢查的文檔、如何測試它。將整個計劃作為小任務交給他們。乾燥。亞格尼。時分驅動。頻繁提交。
假設他們是一位熟練的開發人員,但對我們的工具集或問題領域幾乎一無所知。假設他們不太瞭解良好的測試設計。
開始時宣佈:“我正在使用寫作計劃技能來創建實施計劃。”
上下文: 這應該在專用工作樹中運作(透過腦力激盪技能創建)。
將計劃保存到:
docs/plans/YYYY-MM-DD-<feature-name>.mdWrite a comprehensive implementation plan assuming engineers have zero background in our codebase and questionable taste. Document everything they need to know: which files, code, tests to touch for each task, documentation they might need to check, how to test it. Break the entire plan into small tasks for them. DRY. YAGNI. Time-boxed. Frequent commits.
Assume they are a skilled developer but know almost nothing about our toolset or problem domain. Assume they don't understand good test design very well.
Announce at the start: "I'm using the Writing Plan skill to create an implementation plan."
Context: This should operate in a dedicated worktree (created via the Brainstorming skill).
Save the plan to:
docs/plans/YYYY-MM-DD-<feature-name>.md一口大小的任務粒度
Bite-sized Task Granularity
每一步都是一個動作(2-5 分鐘):
- “編寫失敗的測試”-步驟
- 「運行它以確保它失敗」-步驟
- “實現最少的程式碼以使測試通過” - 步驟
- “運行測試並確保它們通過”- 步驟
- “提交”-步驟
Each step is an action (2-5 minutes):
- "Write failing test" step
- "Run it to ensure it fails" step
- "Implement minimal code to make the test pass" step
- "Run tests and ensure they pass" step
- "Commit" step
計劃文檔標題
Plan Document Title
每個計劃必須以此標題開頭:
markdown
undefinedEvery plan must start with this title:
markdown
undefined[Feature Name] Implementation Plan
[Feature Name] Implementation Plan
For Claude: REQUIRED SUB-SKILL: Use superpowers:executing-plans to implement this plan task-by-task.
Goal: [One sentence describing what this builds]
Architecture: [2-3 sentences about approach]
Tech Stack: [Key technologies/libraries]
undefinedFor Claude: REQUIRED SUB-SKILL: Use superpowers:executing-plans to implement this plan task-by-task.
Goal: [One sentence describing what this builds]
Architecture: [2-3 sentences about approach]
Tech Stack: [Key technologies/libraries]
undefined任務結構
Task Structure
markdown
undefinedmarkdown
undefinedTask N: [Component Name]
Task N: [Component Name]
Files:
- Create:
exact/path/to/file.py - Modify:
exact/path/to/existing.py:123-145 - Test:
tests/exact/path/to/test.py
Step 1: Write the failing test
python
def test_specific_behavior():
結果 = 函數(輸入)
斷言結果==預期Step 2: Run test to verify it fails
Run:
Expected: FAIL with "function not defined"
pytest tests/path/test.py::test_name -vStep 3: Write minimal implementation
python
def函數(輸入):
預期回報Step 4: Run test to verify it passes
Run:
Expected: PASS
pytest tests/path/test.py::test_name -vStep 5: Commit
bash
git add tests/path/test.py src/path/file.py
git commit -m“壯舉:新增特定功能”undefinedFiles:
- Create:
exact/path/to/file.py - Modify:
exact/path/to/existing.py:123-145 - Test:
tests/exact/path/to/test.py
Step 1: Write the failing test
python
def test_specific_behavior():
result = function(input)
assert result == expectedStep 2: Run test to verify it fails
Run:
Expected: FAIL with "function not defined"
pytest tests/path/test.py::test_name -vStep 3: Write minimal implementation
python
def function(input):
return expectedStep 4: Run test to verify it passes
Run:
Expected: PASS
pytest tests/path/test.py::test_name -vStep 5: Commit
bash
git add tests/path/test.py src/path/file.py
git commit -m "Feat: Add specific feature"undefined記住
Remember
- 始終精確的文件路徑
- 計劃中的完整代碼(不是“添加驗證”)
- 具有預期輸出的精確命令
- 用@語法參考相關技巧
- DRY、YAGNI、TDD、間隙空間
- Always use exact file paths
- Complete code in the plan (not "add validation")
- Exact commands with expected output
- Reference related skills using @ syntax
- DRY, YAGNI, TDD, time-boxed
執行交接
Execution Handoff
保存計劃後,提供執行選擇:
“計劃完成並保存到。兩種執行選項:
docs/plans/<filename>.md1.子代理驅動(本次會議) - 我為每個任務分配新的子代理,在任務之間進行審查,快速迭代
2.並行會話(單獨) - 使用執行計劃打開新會話,使用檢查點批量執行
哪一種方法? ”
如果選擇子代理驅動:
- 所需的子技能: 使用超能力:子代理驅動開發
- 留在本次會議
- 每個任務新鮮的子代理+程式碼審查
如果選擇並行會議:
- 引導他們在工作樹中打開新會話
- 所需的子技能: 新會話使用超能力:執行計劃
After saving the plan, provide execution options:
"Plan completed and saved to . Two execution options:
docs/plans/<filename>.md1. Sub-agent Driven (This Session) - I assign new sub-agents for each task, conduct reviews between tasks, iterate quickly
2. Parallel Session (Separate) - Open a new session using Execute Plan, batch execute using checkpoints
Which approach? "
If Sub-agent Driven is selected:
- Required Sub-skill: Use superpowers:sub-agent-driven-development
- Stay in this session
- Fresh sub-agent + code review for each task
If Parallel Session is selected:
- Guide them to open a new session in the worktree
- Required Sub-skill: New session uses superpowers:executing-plans