OpenWriterControle/.opencode/context/core/workflows/component-planning.md
Darren Hinde 9a58a03284
fix(context): Add frontmatter to context files with missing metadata (#97)
* feat(context): Add project intelligence system

- Add project-intelligence.md standard (what & why)
- Add project-intelligence-management.md guide (how to manage)
- Add project-intelligence/ folder with 6 context files:
  - navigation.md (quick overview & routes)
  - business-domain.md (business context, users, value)
  - technical-domain.md (stack, architecture, decisions)
  - business-tech-bridge.md (business → technical mapping)
  - decisions-log.md (decision history with context)
  - living-notes.md (active issues, debt, questions)
- Update standards navigation with new entries
- Register all new contexts in registry.json

This enables projects to document and maintain business + technical
context for quick onboarding and lasting project understanding.

* fix(registry): Add project-intelligence to profiles

- Add context:project-intelligence/* to developer, business, full, and advanced profiles
- Ensures project intelligence contexts are included when these profiles are selected

This complements commit c731328 which added the context entries.

* docs(core): refine project intelligence system and deprecate legacy context

* feat(claude): simplify Claude Code integration with unified installer

- Create single install-claude.sh script that auto-converts and installs
- Remove redundant bootstrap-install.sh and install.sh scripts
- Delete stale plugin/ directory (now auto-generated from .opencode/)
- Add Claude Code documentation (agent-skills, create-subagents, hooks, plugins)
- Add installation profiles (essential, developer, business, full, advanced)
- Update README with Claude Code beta feature documentation
- Add component-planning.md workflow context
- Clean up demo files (assets/, src/)
- Update registry with new context entries
- Add gitignore rule for converter/generated/

The new installer:
- Converts agents automatically from .opencode/agent/
- Installs fresh converted agents to ~/.claude/plugins/
- Single command: ./integrations/claude-code/install-claude.sh

* fix: update README to use new install-claude.sh script

The Claude Code integration now uses the unified install-claude.sh script
instead of the removed bootstrap-install.sh and install.sh scripts.

* fix(context): Add frontmatter to context files with missing metadata

- Add YAML frontmatter with description to system-builder-templates
- Add frontmatter to ui/web/design context files
- Fixes validation warnings in validate-component.sh

* fix(registry): Remove orphaned context entries for deleted files

- Remove copywriting-frameworks entry (file moved to content-creation/principles/)
- Remove tone-voice entry (file moved to content-creation/principles/)
2026-01-17 21:29:48 +00:00

3.3 KiB

Component-Based Planning Workflow

Overview

This workflow replaces "Monolithic Planning" (planning everything at once) with "Iterative Component Planning". It is designed for complex features that require breaking down into functional units.

Core Philosophy

"Plan the System, Build the Component." Don't try to write a detailed plan for the entire system upfront. Create a high-level roadmap, then zoom in to plan one component in detail before executing it.

The Two-Level Plan Structure

Level 1: The Master Plan (Roadmap)

File: .tmp/sessions/{id}/master-plan.md Purpose: High-level architecture and dependency graph. Content:

  • System Architecture Diagram (ASCII)
  • List of Components (e.g., Auth, Database, API, UI)
  • Dependency Order (What must be built first?)
  • Global Standards/Decisions

Level 2: The Component Plan (Active Spec)

File: .tmp/sessions/{id}/component-{name}.md Purpose: Detailed execution steps for the current focus. Content:

  • Interface Definition: Types, function signatures, API contracts.
  • Test Strategy: What specific cases will be tested?
  • Task List: Atomic steps (Create file, Write test, Implement logic).
  • Verification: How do we know this component is done?

Workflow Steps

Phase 1: System Design (The Master Plan)

  1. Analyze: Understand the full feature request.
  2. Decompose: Break the system into functional Components (e.g., "User Service", "Email Worker", "Frontend Form").
  3. Draft Master Plan: Create master-plan.md.
  4. Approve: Get user buy-in on the architecture and order.

Phase 2: Component Execution Loop

Repeat this for each component in the Master Plan:

  1. Select Component: Pick the next unblocked component.
  2. Draft Component Plan: Create component-{name}.md.
    • Define the exact interface/types first.
    • List the atomic implementation steps.
  3. Approve: Show the detailed component plan to the user.
  4. Execute:
    • Load component-{name}.md into TodoWrite.
    • Implement -> Validate -> Check off.
  5. Integrate: Update master-plan.md to mark component as complete.

When to Use This

  • Complex Features: >3 files, multiple layers (DB + API + UI).
  • Unknowns: When later parts of the system depend on earlier decisions.
  • Large Scope: Anything taking >2 hours.

Example Master Plan (master-plan.md)

# Master Plan: E-Commerce Checkout

## Architecture
[Cart] -> [Order Service] -> [Payment Gateway]
                       -> [Inventory Service]

## Component Order
1. [ ] **Inventory Service** (Check stock)
2. [ ] **Order Service** (Create order record)
3. [ ] **Payment Integration** (Stripe)
4. [ ] **Checkout UI** (React components)

Example Component Plan (component-inventory.md)

# Component: Inventory Service

## Interface
```typescript
interface InventoryManager {
  checkStock(sku: string): Promise<boolean>;
  reserve(sku: string, quantity: number): Promise<void>;
}

Tasks

  • Define InventoryManager interface in src/types.ts
  • Create mock implementation for tests
  • Implement checkStock logic with DB query
  • Add unit tests for race conditions