Skip to content
Agent Orchestration
Skill

/handover

Write the README and the first page: where the app is, how to run and publish it, and how to back up its data.

BOOST
From plugin
mulmoterminal
24574 skills
Install
$ npx -y skills add receptron/mulmoterminal --skill handover --agent claude-code

How it fires

How this skill gets triggered: by you, by Claude, or both.

  • Fires itselfAuto-invocation. Claude auto-loads it when your prompt matches the work.Auto-invocation is when the right skill fires by itself at the right moment, driven by a FLOW.md router and a hook, instead of you invoking it by name. It is the difference between a skill being installed and a skill actually getting used.Read the full definition →
  • You can call itInvoke it directly when you want it.
  • Slash command/handover

Context preview

The summary Claude sees to decide when to auto-load this skill.

Write the README and the first page: where the app is, how to run and publish it, and how to back up its data.

SKILL.md

handover.SKILL.md
name: blueprint-cloudflare-handover
description: "Write the README and the first page: where the app is, how to run and publish it, and how to back up its data."

Handover

Write `README.md` in the spec's language, for someone who is not an engineer:

1. What the app does, in two sentences, and where it is published (the URL in `.blueprint/deploy-url`). 2. Running it on this computer: `yarn install` once, then `yarn build` and `yarn start`, then the address to open. 3. Publishing a change: try it with `yarn start` first, then `yarn deploy` (it runs `wrangler deploy`); a schema change is a new migration, applied with `yarn wrangler d1 migrations apply DB --remote` before the deploy. 4. The data: it is in D1 on Cloudflare. Back it up with `yarn wrangler d1 export DB --remote --output backup.sql`, and how to restore it. 5. Who can open it (anyone with the URL, or only signed-in people) and how to change that.

Then write `.blueprint/start-here.md`, in the spec's language — what the person sees the moment the build finishes:

1. **Open it** — the published URL, and who can sign in and how. 2. **Check that it works** — one checkbox per must-have in the spec, each an action and what should happen (`- [ ] 本を 1 冊登録する → 一覧に出る`), with the app's real screen and button names; one for someone who should be refused being refused. 3. **The next change** — try it with `yarn start`, then `yarn deploy`. 4. **Your data** — the one-line backup, and where the security review is (`.blueprint/security-review.md`).

Done when the check passes: the README explains `yarn start`, `yarn deploy` and `wrangler d1 export`, and `.blueprint/start-here.md` names the published URL and has the checklist.

Always

  • Read `.blueprint/spec.md` first. It is the agreed specification; do not widen it.
  • When you need a decision, ask it through the blueprint question tool and stop. Do not guess.
  • Do not run `git init`: a new repository loses the folder's trust and the next unattended step stops at Claude

Code's trust prompt. The user adds git themselves after the build if they want it.

  • Say you are done by stopping; the executor runs the check. Do not claim success yourself.
Read more
Ships withmulmoterminal

Run multiple coding-agent sessions in parallel — and see which one needs you. A browser terminal for parallel AI coding agents: several sessions side by side, each in its own cell, with the one that needs you marked in colour.

Get the whole plugin
Stats
255
Stars
32
Forks
Active
Maintenance
TypeScript
Language
MIT
License
4h ago
Last commit
4mo ago
Created
18h ago
Added

Repo: receptron/mulmoterminal

Other skills on mulmoterminal.