From clippy
Internal skill — decomposes design artifacts into a task graph with units, dependencies, and parallelism groups. Called by autopilot after design is complete.
How this skill is triggered — by the user, by Claude, or both
Slash command
/clippy:decomposeThe summary Claude sees in its skill listing — used to decide when to auto-load this skill
Reads design artifacts from `.ai/design/`, produces a task graph for
Reads design artifacts from .ai/design/, produces a task graph for
autonomous implementation.
Read .ai/design/DESIGN.md, .ai/design/STACK.md, and all files in
.ai/design/contracts/. Understand components, connections, tech stack,
and testing approach.
Each component/service/module is a candidate unit. A unit should be:
depends_on: which units must complete firstparallel_group: units with same dependencies and no inter-dependency
get the same group nameFor each unit:
contracts_read: contract files to read before implementingcontracts_produced: files this unit creates that others need.ai/design/contracts/.
If missing: STOP and report the gap. Do not guess..ai/task-graph.yaml:
version: 1
units:
- id: unit-001
name: "..."
description: "..."
depends_on: []
contracts_read: []
contracts_produced: []
parallel_group: null
.ai/state.yaml:
version: 1
status: pending
current_unit: null
units_completed: []
units_failed: []
units_remaining: [unit-001, ...]
consecutive_failures: 0
error: null
.ai/constraints.yaml — preserve if exists (may have design-phase
constraints). Create empty if not:
version: 1
constraints: []
.ai/checkpoints/ — create empty directory.
Validate: re-read task-graph.yaml, verify all fields, verify all contracts_read files exist on disk, verify no circular dependencies.
Present task graph summary: N units, M parallel groups, dependency chain depth, any concerns about sizing. Recommend proceeding.
🔚 DECOMPOSE PHASE COMPLETE
Units: N
Parallel groups: M
Dependency depth: D
State written to: .ai/
npx claudepluginhub gunther-schulz/coding-clippyDecomposes PRDs and Tech Specs into parallel-executable tasks with contracts, prompts, and dependency graphs for multi-agent execution.
Decompose projects or tracks into modules with dependency mapping. Updates architecture.md and generates HLD/LLD artifacts to drive implement, deploy-checklist, and upload sign-off workflows.
Decomposes plans into file-dependency tasks, groups into parallel Waves, outputs Mermaid graphs and split TODO directories. Use to parallelize implementation and visualize deps.