sdd-init

sdd-init

Purpose

Create a new spec directory for a feature, following the SDD naming convention and populating it with template files. This is the entry point to the SDD workflow.

When to Use

• Starting work on a new feature that needs a spec
• Beginning a significant refactor that should be documented
• Planning an architectural change that affects multiple packages

When NOT to Use

• Bug fixes, typo corrections, small refactors (no spec needed)
• A spec already exists for the feature (use /sdd-spec to refine it)

Step-by-Step Instructions

1. Read specs/README.md to get the next available KNOW number from the "Next Available Number" section

2. Ask the user (if not provided as argument):

   • Short title for the feature (will become directory name in kebab-case)
   • One-line summary of what the feature does

3. Create the spec directory:

   specs/KNOW-{NNN}-{kebab-case-title}/
   

4. Copy and populate templates from specs/_templates/:

   • requirements.template.md → requirements.md (replace {TITLE} and {NNN})
   • design.template.md → design.md (replace {TITLE} and {NNN})
   • tasks.template.md → tasks.md (replace {TITLE} and {NNN})

5. Update specs/README.md:

   • Increment the "Next Available Number"
   • Add an entry to the "Spec Index" table with status draft

6. Report to the user:

   • Spec directory path
   • Suggest next step: /sdd-spec KNOW-{NNN} to write the spec content

Arguments

• $ARGUMENTS (optional): Feature title. If not provided, ask the user.

Example

User: /sdd-init push-based capture

Result:

Created: specs/KNOW-001-push-based-capture/
  ├── requirements.md (template)
  ├── design.md (template)
  └── tasks.md (template)

Updated: specs/README.md (index + next number)

Next step: /sdd-spec KNOW-001

Reference Files

• specs/README.md — Naming conventions, next available number
• specs/_templates/ — Template files to copy
• docs/workflows/sdd-overview.md — Full SDD workflow context