Execute the full P1 planning workflow — parallel codebase exploration, domain decomposition, slice authoring, and artifact generation — before any implementation begins. Load this skill whenever the user says "plan this feature", "create slices", "plan feature X", "I want to plan before implementing", or any request that needs decomposition into independently mergeable slices. Use proactively for any non-trivial feature (more than 3 files). Do not wait for the user to explicitly say "P1" — trigger on planning intent.
This skill executes the P1 planning phase: parallel exploration, vertical slice decomposition, and artifact generation. It produces the slice specs, plan.md, and queue.md that the downstream P2-P9 workflow consumes. No source code is changed during this skill.
Read these files now:
.claude/skills/slice-authoring/SKILL.md — every slice you write must pass its Implementation-Readiness Checklist.claude/docs/PROJECT.md — stack context and Repo Registry (workspace mode).claude/docs/BRANCHING.md — feature branch naming conventionsSpawn 2–3 Explore agents in the same message (never sequentially, never inline). Agents cover more ground and preserve your context window for decomposition work.
Assign each agent a distinct focus:
Agent 1 — Existing implementations: Find code already related to the feature: similar features, reusable utilities, relevant services/models/components. Record file paths and sizes (wc -l). The goal is to identify what can be reused — avoid proposing new code when suitable implementations exist.
Agent 2 — Consumer graph: For each file likely to change, find all files that import or depend on it. Categorize