# Plan And Stop

> Inspect and plan a software task, then stop before editing. Use for planning, scoping, implementation design, or review checkpoints. Do not use when the user requests immediate implementation.

- Skill: `davidbasilefilho/plan-and-stop` (Agent Skill, multi-file: 2 files)
- Install (CLI): `npx skillmds@latest add davidbasilefilho/plan-and-stop`
- Raw SKILL.md: https://api.skillmd.com/api/skills/davidbasilefilho/plan-and-stop/raw
- Safety review: pending
- Works with: Claude Code, Claude.ai, OpenAI Codex
- Category: Productivity
- Author: davidbasilefilho (https://skillmd.com/u/davidbasilefilho)
- Updated: 2026-09-21
- Page: https://skillmd.com/skills/davidbasilefilho/plan-and-stop

---


# Plan and Stop

Load `caveman` and use it to write the plan minimally.

1. Read applicable `AGENTS.md` and repo instructions.
2. Inspect only code, config, docs, and tests needed for task understanding.
3. Do not edit implementation files, spawn subagents, run destructive commands, commit, push, or open PRs.
4. State minor assumptions; record material uncertainty.
5. Produce concrete plan with outcome, current behavior/evidence, likely files/components, ordered steps, validation/tests, risks, assumptions, and open decisions.
6. Create `.agents/plans/` if needed. Save plan as `.agents/plans/<timestamp>-<task>.md`.
7. Stop. Report plan path. Tell user to load `review-plan` next. Never implement in same turn.