Writing Plans

Overview

Write comprehensive implementation plans assuming the engineer has zero context for our codebase and questionable taste. Document everything they need to know: which files to touch for each task, code, testing, docs they might need to check, how to test it. Give them the whole plan as bite-sized tasks. DRY. YAGNI. TDD. Frequent commits.

Assume they are a skilled developer, but know almost nothing about our toolset or problem domain. Assume they don't know good test design very well.

Announce at start: "I'm using the writing-plans skill to create the implementation plan." Save plans to: [Save implementation plan to the epic/user-story folder. If an implement plan already exists for the epic or user story, use that file by overwriting the placeholder text. If no plan exists, create a new one. Display location to user]

Required Pre-Flight Checklist

1. •Make sure test-driven-development skill %% force-extract %% is in your context window
2. •Make sure sub-agent-driven-development %% force-extract %% is in your contenxt window

Bite-Sized Task Granularity

Each step is one action (2-5 minutes):

• •"Write the failing test" - step
• •"Run it to make sure it fails" - step
• •"Implement the minimal code to make the test pass" - step
• •"Run the tests and make sure they pass" - step
• •"Commit" - step

Task Types

Label each task with its type. Different types need different structures.

Infrastructure Tasks

When: Directory setup, config files, .gitignore changes Structure:

## Task N — [Name]

**Type:** Infrastructure

### Files
- `.path/to/dir/` (CREATE directory)
- `.path/to/config.json` (CREATE)

### Step 1: Create structure
[commands]

### Step 2: Verify
[verification command with expected output]

### Step 3: Commit

TDD Tasks

When: Any code that can be tested Key rule: Explicit "expect FAIL" and "expect PASS" markers

## Task N — [Name]

**Type:** TDD

### Files
- `path/to/code.sh` (CREATE)
- `path/to/test.sh` (CREATE & TEST)

### Step 1: Write the failing test
[complete test code]

### Step 2: Run test — expect FAIL
Run: `[command]`
**Expected:** FAIL — [specific reason why it fails]

### Step 3: Write minimal implementation
[complete implementation code]

### Step 4: Run test — expect PASS
Run: `[command]`
**Expected:** PASS

### Step 5: Commit

For bash scripts: Consider splitting into 2 tasks:

• •Task N (RED): Test harness that expects script to exist → fails
• •Task N+1 (GREEN): Implementation that makes test pass

Validation Tasks

When: Performance benchmarks, profiling Structure:

## Task N — [Name]

**Type:** Validation

### Step 1: Create benchmark script
[script code]

### Step 2: Run and document results
Run: `[command]`
**Expected:** [target metrics]

### Step 3: Commit

Integration Tasks

When: settings.json changes, end-to-end verification Structure:

## Task N — [Name]

**Type:** Integration

### Step 1: Create integration test
[test code]

### Step 2: Run — expect FAIL
**Expected:** FAIL — [config not yet updated]

### Step 3: Update config/settings
[changes to make]

### Step 4: Run — expect PASS
**Expected:** PASS

### Step 5: Commit

Plan Document Header

Every plan MUST start with this header:

# [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]

---

Task Structure (TDD Example)

## Task {{task-number}} - {{task-name}}

### Files
- `exact/path/to/file.py` (CREATE)
- `exact/path/to/existing.py:123-145` (MODIFY)
- `tests/exact/path/to/test.py` (CREATE & TEST)

### Step 1: Write the failing test

```python
def test_specific_behavior():
    result = function(input)
    assert result == expected
```

### Step 2: Run test to verify it fails

Run: `pytest tests/path/test.py::test_name -v`
Expected: FAIL with "function not defined"

### Step 3: Write minimal implementation

```python
def function(input):
    return expected
```

### Step 4: Run test to verify it passes

Run: `pytest tests/path/test.py::test_name -v`
Expected: PASS

### Step 5: Commit
Use `create-git-commit` skill to commit

Remember

• •Exact file paths always
• •Complete code in plan (not "add validation")
• •Exact commands with expected output
• •Reference relevant skills with @ syntax
• •DRY, YAGNI, TDD, frequent commits

Execution Handoff

After saving the plan, offer execution choice:

"Plan complete and saved to {{implement-plan-path}}. Ready to execute?

1. Set up implementation worktree - Create isolated workspace for this implementation

2. Execute in this session (Subagent-Driven) - I dispatch fresh subagent per task, review between tasks, fast iteration

3. Execute in parallel session - Open new session with executing-plans, batch execution with checkpoints

If Worktree Setup chosen:

• •REQUIRED SUB-SKILL: Use git-using-worktrees
• •After worktree created, offer execution options again:
  • •Execute now in this session (Subagent-Driven)
  • •Execute later in new session (Parallel Session)

If Subagent-Driven chosen:

• •REQUIRED SUB-SKILL: Use subagent-driven-development
• •Stay in this session
• •Fresh subagent per task + code review

If Parallel Session chosen:

• •Guide them to open new session in worktree
• •REQUIRED SUB-SKILL: New session uses executing-plans