# Warm Feature Flow

> Warm the agent, reset git to a clean base branch, then work and create a draft PR. Use when this capability is needed.

- Skill: `tomevault-io/warm-feature-flow` (Agent Skill, multi-file: 2 files)
- Install (CLI): `npx skillmds@latest add tomevault-io/warm-feature-flow`
- Raw SKILL.md: https://api.skillmd.com/api/skills/tomevault-io/warm-feature-flow/raw
- Safety review: pending (external: skill-scanner PASS, skillspector PASS)
- Works with: Claude Code, Claude.ai, OpenAI Codex
- Category: AI & ML
- Author: tomevault-io (https://skillmd.com/u/tomevault-io)
- Updated: 2026-09-17
- Page: https://skillmd.com/skills/tomevault-io/warm-feature-flow

---


# Warm Feature Flow

Use this at the start of a new feature to ensure the repo is clean, the agent is warmed, and the draft PR flow is ready at the end.

## Steps

1. **VCS check**
   - If `.jj/` exists, use `jj`; otherwise use `git`.
   - Run `status` and confirm there are no unrelated changes. If there are, stop and ask the user what to do.

2. **Reset to base branch**
   - Checkout base branch (`dev` for normal development, `master` only for releases).
   - Pull latest changes.

3. **Create a fresh feature branch**
   - Branch name format: `feat/<short-scope>` or `fix/<short-scope>`.

4. **Warm the agent**
   - Run the warm-agent workflow (Task tool, subagent `warm-agent`).
   - Focus on current feature area only; avoid unrelated exploration.

5. **Implement the feature**
   - Make changes, run relevant tests if needed.

6. **Create draft PR**
   - Run the `create-draft-pr` skill (or `gh pr create --draft` if needed).

## Notes

- If the base branch is behind, always pull before branching.
- If you must keep local changes, create a separate branch or stash before continuing.

---
> Converted and distributed by [TomeVault](https://tomevault.io/claim/noisemaker111) — claim your Tome and manage your conversions.
<!-- tomevault:4.0:skill_md:2026-04-16 -->