Brain Dump → Backlog Skill

Purpose

Transform unstructured thoughts into organized, actionable backlog tasks.

Workflow

Step 1: Capture the Brain Dump

When the user brain-dumps, capture ALL items. Don't filter. Examples:

• •"We need to fix the login bug, add dark mode, and refactor the API"
• •"I was thinking about: 1) better onboarding 2) performance optimization 3) new dashboard"
• •Random list of features, bugs, ideas, tech debt

Step 2: Categorize Each Item

For each item, determine:

• •Type: feature | bug | tech-debt | research | polish | infrastructure
• •Priority: critical | high | medium | low
• •Area: Which label(s) apply? (effects, 3d-engine, launch, architecture, ai-bundling, games, economy, multiplayer, etc.)

Step 3: Create Tasks

Use backlog task create for each item:

backlog task create "Title" --desc "Description" --status "To Do" --priority {critical|high|medium|low} -l {label1} -l {label2}

Step 4: Set Dependencies (if obvious)

If items clearly depend on each other:

backlog task edit {task-id} --dep {dependency-task-id}

Step 5: Report Back

Show the user:

1. •How many tasks were created
2. •The organized board view
3. •Any dependencies that were set

Label Categories

Use these labels consistently:

Example Brain Dump Response

User: "Okay so we need to: fix that crash in the shader compiler, add the new blur effect, make sure the 3D camera works on web, and oh yeah we should probably document the scripting API"

Response:

Created 4 tasks from your brain dump:

1. **TASK-11** - Fix crash in shader compiler (critical, effects)
2. **TASK-12** - Add new blur effect (high, effects)
3. **TASK-13** - 3D camera web compatibility (high, 3d-engine)
4. **TASK-14** - Document scripting API (medium, docs)

Board updated! View with: backlog board

Tips

• •Don't overthink categorization — use best judgment
• •When in doubt, ask: "Should [item] be high or medium priority?"
• •Group related items under the same label
• •If >10 items, suggest creating a milestone