SDD Implementation Skill

Build what has been planned, following specs and todo-lists precisely.

When to Use

• Executing planned implementations
• Code generation from specifications
• Building features according to plan

Protocol

Step 1: Load the Plan

Read: plan.md → spec.md → tasks.md → todo-list.md

Step 2: Execute Systematically

1. Read entire list before starting
2. Execute in order — respect dependency chains
3. Mark completion — - [ ] → - [x] immediately
4. Document blockers — never skip silently, use [BLOCKED: reason]

Step 3: Follow Patterns

Reference references/patterns.md for project conventions and implementation patterns.

Step 4: Track Progress

Use scripts/progress.sh to visualize completion status.

Step 5: Report

## Implementation Summary

### Completed
- [x] Task 1: description

### Files Created/Modified
- `path/to/file.ts`: [purpose]

### Blockers Encountered
- [blocker and resolution]

### Discoveries
- [anything that should update specs]

Anti-Patterns

• Skipping tasks without explanation
• Marking items done without completing them
• Implementing differently than planned without noting why
• Ignoring blockers instead of documenting them

Integration

• After completion, sdd-verifier subagent validates work (spawned as a child subagent)
• Recall/persist project knowledge via the sdd-memory skill (no-op when the memory provider is standard)
• Discoveries trigger sdd-evolve skill for spec updates
• Use the ask question tool for ambiguous requirements