Skip to content
just-say-so

/just-say-so:init

What it does

Section titled “What it does”

Sets a project up in four steps and prints each one. It writes a .vale.ini when the project has none, with StylesPath = styles, Vocab = Project and an empty accept.txt, the JustSaySo package pinned at the plugin’s version, and [*] BasedOnStyles = JustSaySo. When a .vale.ini exists, it pins the package and leaves the rest alone. It runs vale sync. It writes .just-say-so.json for the flags you gave, merged over an existing file. It wires the commit gate for the hook tooling the repo already has. The doctor report follows.

Synopsis

Section titled “Synopsis”
/just-say-so:init [--files MODE] [--publish MODE] [--gh] [--mcp PATTERN]... [--reply MODE] [--hook auto|pre-commit|husky|git|none] [--dry-run]

Arguments

Section titled “Arguments”
Argument Meaning
--files MODE, --publish MODE, --reply MODE off, warn, or block for that target. Omitted targets keep their defaults.
--gh Check the gh commands that publish prose.
--mcp PATTERN An MCP tool pattern to check, such as mcp__atlassian__*. Repeat for more.
--hook auto|pre-commit|husky|git|none Which commit hook to wire. auto, the default, picks the pre-commit framework when .pre-commit-config.yaml exists, Husky when .husky/ exists, a plain .git/hooks/pre-commit when the repo has a package.json, and otherwise prints the pre-commit snippet.
--dry-run Print every step without writing.

With no arguments, the skill asks one question first: which targets you want checked beyond the defaults, and whether the agent publishes through gh or an MCP tool.

What it reads and writes

Section titled “What it reads and writes”

Reads the project root (the nearest .git), any existing .vale.ini and .just-say-so.json, and the repo’s hook tooling. Writes .vale.ini, styles/config/vocabularies/Project/accept.txt, .just-say-so.json, and the hook file for the tooling it chose. Runs vale sync, which downloads the package into styles/.

What it prints

Section titled “What it prints”

One line per step, prefixed just-say-so init:, then the full doctor report. When a step needs a command only you can run, such as pre-commit install or a dev dependency install, the skill repeats that command on its own line at the end.

Example

Section titled “Example”

A bare repository, dry run:

just-say-so init: project root: /Users/you/proj (dry run, nothing written)
just-say-so init: would write /Users/you/proj/.vale.ini with the JustSaySo package, StylesPath = styles, and Vocab = Project
just-say-so init: would create /Users/you/proj/styles/config/vocabularies/Project/accept.txt
just-say-so init: would run vale sync
just-say-so init: no plugin settings given, so the defaults apply (files warn, publish block with no source, reply warn)
just-say-so init: commit gate: add this to .pre-commit-config.yaml (the file does not exist yet) and run pre-commit install:
just-say-so init:     repos:
just-say-so init:       - repo: https://github.com/ifandelse/just-say-so
just-say-so init:         rev: v0.18.1
just-say-so init:         hooks:
just-say-so init:           - id: vale-staged