/planning Skill
Taxonomy note: This skill provides domain expertise (HOW) for decomposing work under uncertainty. It is the primary skill for the [[decompose]] workflow. See [[TAXONOMY.md]] for the skill/workflow distinction.
Patterns for decomposing work under genuine uncertainty.
Purpose
Provides guidance for breaking down complex goals when the path forward is unknown. Use when standard decomposition feels premature or when you need to discover what you don't know.
Coordination with Workflows
This skill is the HOW for the [[decompose]] workflow's steps. When decomposing an epic that has an assigned workflow, use the workflow's steps as the decomposition skeleton — then apply the patterns below to handle uncertainty within each step. See [[decomposition-patterns]] for workflow-step mapping examples.
When to Use
- Multi-month projects with unclear dependencies
- "What does X actually require?" questions
- Creating task hierarchies that handle uncertainty
- Deciding between spike tasks vs placeholder tasks
- Structuring knowledge flow between related tasks
Key Patterns
See reference documents for detailed guidance:
- [[spike-patterns]] - When to investigate vs commit to implementation
- [[dependency-types]] - Hard vs soft dependency decisions
- [[knowledge-flow]] - Propagating findings between sibling tasks
Core Principle
We decompose to discover what we don't know, not because we know what to do.
Quick Reference
Spike vs Placeholder
| Situation | Use Spike | Use Placeholder |
|---|---|---|
| "We don't know if X is possible" | Investigate first | |
| "We know X is needed, details TBD" | Capture intent | |
| "Implementation approach unclear" | Prototype/probe |
Dependency Type Selection
Decision heuristic: "What happens if the dependency never completes?"
- Impossible or wrong output → Hard dependency (
depends_on) - Still valid but less informed → Soft dependency (
soft_depends_on)
Related
- [[decompose]] workflow - Overall decomposition process
- [[base-task-tracking]] - Task lifecycle management