survey
Measure what chaff reports on the named documents today, as the kind the person chose, and propose the setup — changing nothing…
Set chaff up in the folder: chaff.yaml for the chosen genre, today's findings shelved in .chaff-baseline.json, and, when asked, a GitHub workflow that puts new findings on the pull request's lines.
$ npx -y skills add receptron/mulmoterminal --skill apply --agent claude-codeHow it fires
How this skill gets triggered: by you, by Claude, or both.
/applyContext preview
The summary Claude sees to decide when to auto-load this skill.
Set chaff up in the folder: chaff.yaml for the chosen genre, today's findings shelved in .chaff-baseline.json, and, when asked, a GitHub workflow that puts new findings on the pull request's lines.
name: blueprint-adopt-apply description: "Set chaff up in the folder: chaff.yaml for the chosen genre, today's findings shelved in .chaff-baseline.json, and, when asked, a GitHub workflow that puts new findings on the pull request's lines."
The person approved the setup in `.blueprint/adopt.json`. The documents stay exactly as they are: fixing what is shelved is 「文書を整える」's work, later.
Write the folder's `chaff.yaml` with the genre from `.blueprint/adopt.json` and the documents' `language` (`npx chaffjs init --genre <genre>` writes a commented one). If the folder already has one, set its `genre` to the chosen one and **keep every other line it has** — the check refuses any other line of the old file that is gone. Add the rule levels the person agreed to at the gate, nothing else.
Run `sh <base pack>/checks/chaff.sh baseline <place> …` with every place. It writes `.chaff-baseline.json` in the folder: the findings there today are not reported again; new ones are. Tell the person to commit it with `chaff.yaml`.
Copy `<usecase pack>/templates/chaff.yml` to `.github/workflows/chaff.yml`, replacing `{{PATHS}}` with the places separated by spaces, and change nothing else: the check compares the file with the template filled in that way, line for line. Its `permissions:` let the workflow read the documents (`contents: read`) and upload findings (`security-events: write`), nothing more. A place whose name has a space or a character a shell reads (`$`, `;`, `*`, quotes…) cannot go into the workflow: the interview refuses it, so ask the person to name a folder above it instead. When `ci` is 「作らない」, add nothing under `.github/`.
Write `.blueprint/adopt-report.md` for the person, in their language and in plain words:
workflow `.github/workflows/chaff.yml` when one was made. The check looks for each file's name.
on the pull request's lines); how to lower the shelf later (「文書を整える」, then `chaff baseline` again); that the workflow needs code scanning enabled on the repository (a public repository has it; a private one needs GitHub Advanced Security).
`node <usecase pack>/checks/adopt.mjs apply` and then `node <usecase pack>/checks/report.mjs` (with `BLUEPRINT_BASE` and `BLUEPRINT_USECASE` set) pass.
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.
Measure what chaff reports on the named documents today, as the kind the person chose, and propose the setup — changing nothing…
Answer each question from the named documents only, quoting where the answer is written, or saying plainly that the documents do…
Add the approved answers to FAQ.md when the person chose to keep them, after what it already holds; change nothing otherwise.
Build the spec's operations as the Worker's API, with validation and tests.