← ClaudeAtlas

writing-planslisted

Use when you have an approved spec/PRD/requirements for a multi-step task and need a concrete implementation plan, BEFORE touching code. Produces bite-sized, TDD-structured tasks with exact file paths, real code, and exact verification commands. Triggers after brainstorming/PRD approval or whenever someone says "write a plan", "plan this out", or hands you requirements to build.
StielChancellor/VibeGod-Tech-Team · ★ 0 · AI & Automation · score 62
Install: claude install-skill StielChancellor/VibeGod-Tech-Team
<!-- Adapted from superpowers (https://github.com/obra/superpowers), MIT (c) Jesse Vincent. --> # Writing Plans ## Overview Write comprehensive implementation plans assuming the engineer (a fresh subagent or a future you) has zero context for this codebase. Document everything: which files to touch per task, the actual code, tests, docs to check, how to verify. Give the whole plan as bite-sized tasks. DRY. YAGNI. TDD. Frequent commits. Assume a skilled developer who knows almost nothing about our toolset, problem domain, or good test design. **Announce at start:** "I'm using the writing-plans skill to create the implementation plan." ## Fits in the pipeline This skill drives **Stage 5 — Build Plan (`/build-plan`, foundation-first)**. Input is the approved PRD/spec + module map; output is the coding roadmap + TDD/UAT/smoke/QA plan. **Build the foundation first** (the shared base all modules sit on). End with the ◆ gate: explicitly ask before starting coding agents. Priority: **user > skills > default**; `_shared/vibegod-principles.md` apply (esp. #2 simplicity, #3 surgical, #6 consistency, #8 TDD). **Save plans to:** `docs/vibegod/plans/YYYY-MM-DD-<feature-name>.md` (user preferences override). **Context:** If working in an isolated worktree, it should have been created via the `using-git-worktrees` skill. ## Scope Check If the spec covers multiple independent subsystems, it should have been split during brainstorming. If not, suggest separate plans — one per subsys