Tier 1

decomposition

Break complex goals into simpler, manageable sub-goals

Usage in Claude Code: /decomposition your question here

Decomposition

Overview

Break complex goals into simpler, manageable sub-goals

Steps

Step 1: Assess decomposition need

Determine if and why decomposition is needed:

  • Is goal too big to tackle at once?
  • Are there clearly distinct components?
  • Would parallel work be beneficial?

Step 2: Select decomposition strategy

Choose the most appropriate strategy:

Functional: Use when goal has distinct functions

  • “Build a robot” → locomotion, sensing, control, power

Temporal: Use when goal has natural phases

  • “Launch product” → research, build, test, launch

Structural: Use when goal has physical/logical parts

  • “Build house” → foundation, framing, electrical, plumbing

Stakeholder: Use when different people need different things

  • “Improve satisfaction” → customer needs, employee needs, partner needs

Step 3: Identify sub-goals

Apply selected strategy to identify sub-goals:

  1. List all components/phases/functions
  2. Name each as a sub-goal
  3. Ensure sub-goals cover the whole goal
  4. Check for gaps (is anything missing?)

Sub-goal quality criteria:

  • SIMPLER than parent (can be understood independently)
  • COMPLETE (together achieve the parent)
  • INDEPENDENT (minimal overlap)
  • ACTIONABLE (can be worked on)
  • MEASURABLE (can track progress)

Step 4: Check completeness (MECE)

Verify decomposition is Mutually Exclusive, Collectively Exhaustive:

Mutually Exclusive: Sub-goals don’t overlap

  • If two sub-goals could both address the same work, merge or clarify

Collectively Exhaustive: Sub-goals cover everything

  • If achieving all sub-goals wouldn’t achieve parent, something is missing

Step 5: Identify dependencies

For each pair of sub-goals, determine if:

  • A depends on B (B must complete before A starts)
  • A and B are independent (can happen in parallel)
  • A and B are sequential (must happen in order)

Build dependency graph showing relationships.

Step 6: Calculate critical path

Find the longest chain of dependent sub-goals:

  1. Start from sub-goals with no dependencies
  2. Follow dependency chains
  3. Find longest path (this is your minimum time)

Critical path determines earliest possible completion.

Step 7: Check for further decomposition

For each sub-goal, ask:

  • Is this sub-goal still too complex?
  • Should it be decomposed further?

If yes and within max_depth:

  • Recursively apply decomposition to that sub-goal
  • Note parent-child relationship

When to Use

  • Goal is too complex to tackle directly
  • Need to break down work into phases
  • Multiple parallel efforts needed
  • Goal has multiple distinct components
  • Team needs to divide work
  • Complexity assessment shows “Complicated” or “Complex”

Verification

  • Sub-goals are simpler than original goal
  • Sub-goals together achieve original goal (collective exhaustiveness)
  • Sub-goals don’t overlap significantly (mutual exclusiveness)
  • Dependencies are identified and make sense
  • Critical path is realistic

Input: $ARGUMENTS

Apply this procedure to the input provided.