writing-plans

Imported from andyandsmoothies-ship-it/vtcoon (.agents/skills/writing-plans/SKILL.md). Install upstream with npx skills add andyandsmoothies-ship-it/vtcoon --skill writing-plans. Copyright stays with the author.

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."

Context: If working in an isolated worktree, it should have been created via the superpowers:using-git-worktrees skill at execution time.

Save plans to: docs/plans/improvements/IMP-XXX-<feature-name>_plan.md

• (User preferences for plan location override this default)

Scope Check

If the spec covers multiple independent subsystems, it should have been broken into sub-project specs during brainstorming. If it wasn't, suggest breaking this into separate plans — one per subsystem. Each plan should produce working, testable software on its own.

Pre-Drafting Physical Verification (The 5 Mandatory Checks)

Before writing any task steps or code snippets, you MUST physically inspect the disk using tools:

1. Call-Site Exhaustion (grep_search): Run grep_search on every symbol/function you plan to change across tests/ and src/. Tabulate every caller and every affected test case. Zero unverified assumptions.
2. Subtractive Deletion Range (view_file): Run view_file on target files to inspect exact lines being replaced or deleted. Record exact start/end line numbers and functions to delete. Zero hand-wavy "refactor later".
3. Banned Mechanism Check: Cross-check proposed snippets against project constraints (e.g. anti-programmer-art primitives, bare strings, loose types).
4. Physical Snippet LOC Count Verification: When writing drop-in replacement snippets for tasks, physically count the lines of the snippet (snippet.split('\n').length). If a planned abstraction or proxy exceeds 100 LOC, design it as an isolated standalone file upfront, preventing accidental LOC ceiling breaches.
5. Collection & Adapter Protocol Parity: When specifying a Proxy or Adapter emulating a standard collection (Map, Set, List, Dict), specify all standard protocol methods (CRUD, iteration, size, entries). Never plan partial stubs. For scalar variables, plan simple local single-entry collections (new Map([[k, v]])) with sync-back instead of dynamic proxies.

File Structure

Before defining tasks, use list_dir and grep_search to map out which files will be created or modified and what each one is responsible for. This is where decomposition decisions get locked in.

• Design units with clear boundaries and well-defined interfaces. Each file should have one clear responsibility.
• You reason best about code you can hold in context at once, and your edits are more reliable when files are focused. Prefer smaller, focused files over large ones that do too much.
• Files that change together should live together. Split by responsibility, not by technical layer.
• In existing codebases, follow established patterns. If the codebase uses large files, don't unilaterally restructure - but if a file you're modifying has grown unwieldy, including a split in the plan is reasonable.
• When integrating with external APIs or libraries, use search_web and read_url_content to fetch current documentation before locking in the file structure

This structure informs the task decomposition. Each task should produce self-contained changes that make sense independently.

Task Right-Sizing

A task is the smallest unit that carries its own test cycle and is worth a fresh reviewer's gate. Fold setup, configuration, scaffolding, and documentation steps into the task whose deliverable needs them; split only where a reviewer could meaningfully reject one task while approving its neighbor.

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

Plan Document Header

Every plan MUST start with this header:

# [Feature Name] Implementation Plan

> **For agentic workers:** REQUIRED SUB-SKILL: Use superpowers:subagent-driven-development to implement this plan task-by-task. Steps use checkbox (`- [ ]`) syntax for tracking.

**Goal:** [One sentence describing what this builds]

**Architecture:** [2-3 sentences about approach]

**Architecture Diagram:**

```mermaid
graph TD
    subgraph "Component Name"
        A[Module A] --> B[Module B]
    end

Include a Mermaid diagram showing component relationships and data flow. This diagram should match the architecture description above.

Tech Stack: [Key technologies/libraries]

Spec: [path to the spec/design doc this plan implements — the plan argues from the spec, so the spec travels with it; executors read both]

Global Constraints

[Project-wide requirements — version floors, dependency limits, naming rules, platform requirements — one line each, exact values from the spec.]

System Impact & Blast Radius (3-Way Matrix)

• Risk Dial: [Isolated (Level 1) | Slice-Bound (Level 2) | Systemic/Global (Level 3)]
• Direct Touch: [Files/modules modified]
• Subtractive Audit (Delete/Cleanup): [Obsolete states, listeners, flags, or dead code paths to remove]
• Call-Site Exhaustion: [100% of callers audited via grep_search with a call-site matrix — never rely on default parameters. Tabulate every caller when modifying signatures.]
• Import DAG Check: [Verify upstream imports of modified files to prevent circular dependencies]
• Delta LOC Budget: [For files >= 300 LOC: Current + Delta = Expected; extract submodule if Expected > Ceiling]
• Axis 1 - Downstream Consumers: [Direct callers, UI subscribers, derived stores/caches, event observers]
• Axis 2 - Upstream & Environmental Modifiers: [Active policies, interceptors, feature flags, global modifiers, buffs/debuffs]
• Axis 3 - Exceptional Lifecycle Modes: [Cold start, full state resync/reconnect, session reset, concurrent multi-event mutations, terminal/closed entity states]
• Worst-Case Defense: [Failure mode isolation, fallback guarantees, and blast radius regression tests]

## Task Structure

````markdown
### Task N: [Component Name]

**Files:**
- Create: `exact/path/to/file.py`
- Modify: `exact/path/to/existing.py:123-145`
- Delete: `exact/path/to/obsolete.py` (or obsolete states/listeners to remove)
- Test: `tests/exact/path/to/test.py`

**Interfaces:**
- Consumes: [what this task uses from earlier tasks — exact signatures]
- Produces: [what later tasks rely on — exact names, parameters, return types]

- [ ] **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

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

git add tests/path/test.py src/path/file.py
git commit -m "feat: add specific feature"

## No Placeholders

Every step must contain the actual content an engineer needs. These are **plan failures** — never write them:
- "TBD", "TODO", "implement later", "fill in details"
- "Add appropriate error handling" / "add validation" / "handle edge cases"
- "Write tests for the above" (without actual test code)
- "Similar to Task N" (repeat the code — the engineer may be reading tasks out of order)
- Steps that describe what to do without showing how (code blocks required for code steps)
- References to types, functions, or methods not defined in any task
- Vague quantifiers (e.g. "N test files unaffected", "several callers") without exact file paths or verified grep proof

## Rich Formatting

Use Antigravity's artifact formatting to make plans scannable:

- **File links:** Always use clickable links: `[filename](file:///absolute/path/to/file)` or `[function](file:///path/to/file#L10-L20)`
- **Diff blocks:** Show code changes as diffs when modifying existing files:
  ```diff
  -old_function_name()
  +new_function_name()
   unchanged_line()
  ```
- **GitHub alerts:** Flag critical requirements and breaking changes:
  > [!IMPORTANT]
  > This change requires a database migration

- **Mermaid diagrams:** Use in the architecture section and for complex data flows within tasks

## Self-Review

After writing the complete plan, look at the spec with fresh eyes and check the plan against it. This is a checklist you run yourself — not a subagent dispatch.

**1. Spec coverage:** Skim each section/requirement in the spec. Can you point to a task that implements it? List any gaps.

**2. Placeholder scan:** Search your plan for red flags — any of the patterns from the "No Placeholders" section above. Fix them.

**3. Type consistency:** Do the types, method signatures, and property names you used in later tasks match what you defined in earlier tasks? A function called `clearLayers()` in Task 3 but `clearFullLayers()` in Task 7 is a bug.

If you find issues, fix them inline. No need to re-review — just fix and move on. If you find a spec requirement with no task, add the task.

## Execution Handoff

After saving the plan (using `write_to_file` with `IsArtifact: true`, with the implementation-plan type and `RequestFeedback: true` in `ArtifactMetadata`), confirm execution:

**"Plan complete and saved. Ready to execute with subagent-driven-development?"**

Use `ask_question` to present the confirmation.

User feedback may arrive as inline artifact comments — treat each comment as a change request against that section and confirm resolution in the artifact.

**REQUIRED SUB-SKILL:** Use superpowers:subagent-driven-development
- Fresh subagent per task + two-stage review (spec compliance + code quality)
- Define implementer/spec-reviewer/code-reviewer types upfront