Babysitter task-decomposition
Convert technical plans into actionable development tasks with dependency graphs, effort estimates, and parallelization opportunities.
install
source · Clone the upstream repo
git clone https://github.com/a5c-ai/babysitter
Claude Code · Install into ~/.claude/skills/
T=$(mktemp -d) && git clone --depth=1 https://github.com/a5c-ai/babysitter "$T" && mkdir -p ~/.claude/skills && cp -r "$T/library/methodologies/spec-kit/skills/task-decomposition" ~/.claude/skills/a5c-ai-babysitter-task-decomposition-bda138 && rm -rf "$T"
manifest:
library/methodologies/spec-kit/skills/task-decomposition/SKILL.mdsource content
Task Decomposition
Overview
Decompose a technical plan into actionable development tasks. Each task is atomic, has clear acceptance criteria traceable to specification requirements, and includes dependency information for optimal execution ordering.
When to Use
- After technical plan is approved, before implementation
- When breaking down a large feature into implementable units
- When identifying parallelization opportunities in a task set
- When estimating effort and critical path for a feature
Key Principle
Every task must trace back to at least one specification requirement. Tasks must be atomic enough for a single development session. Dependency graphs must be acyclic.
Process
- Decompose components - Break architecture components into implementable tasks
- Define tasks - Each with id, title, description, acceptance criteria, dependencies, effort
- Build dependency graph - Show task ordering constraints
- Identify parallel groups - Tasks with no mutual dependencies
- Determine critical path - Longest dependency chain
- Generate test tasks - Testing tasks alongside implementation
- Map to requirements - Traceability from tasks to specification
- Human review - Approve task list before implementation
Tool Use
Invoke via babysitter process:
methodologies/spec-kit/spec-kit-planning (task decomposition phase)
Full pipeline: methodologies/spec-kit/spec-kit-orchestrator