Hand the setup to your assistant.
Each of these is a folder with a SKILL.md that a coding
agent reads while working in your repository. They run the Vale CLI, edit your files, and open
your pull request—nothing to sign up for.
Copy the folder into .claude/skills/ in your repository, or
wherever your assistant looks for them. For a setup with no skill support, hand it AGENTS.md instead—it covers the same ground as prose.
Want the engine itself rather than the CLI—scaffolding rules, auditing a style, diffing a rule change across a corpus? That is what Vale CMS and its MCP server are for. These skills stay on the command line.
Install Vale, write a .vale.ini that matches the formats actually present, sync the styles, and get a first run.
Points Vale at the prose rather than the whole repository, and reports the first-run count instead of hiding it behind MinAlertLevel.
Fix alerts a file at a time, applying the replacement each rule defines, one pull request per file.
Error level by default, markup preserved, one branch per file, and never silencing a rule to make the run pass.
Turn a first run on an existing corpus into a decision per rule: fix, downgrade, or switch off.
Reports which few rules produce most of the output, so the answer is a plan rather than a raised alert level.
Add project terms to a vocabulary so spell check accepts them, and commit it as the source it is.
The alternative — disabling Vale.Spelling over one product name — stops checking every other word in the repository.
Run Vale in GitHub Actions, a pre-commit hook, or any other runner, with pinned versions.
Says plainly whether the job is advisory or blocking, since only error sets a non-zero exit code.
Your assistant writes the docs. Something should check them.
A model will write you plausible prose and plausible linter rules with equal confidence. Vale is what tells it which of those actually hold up—in the same turn, not in review.
Setup, fixing alerts, triaging a noisy first run, vocabularies, and CI. Each one runs the CLI in your repository and ends with a report.
Browse the skillsLints each prose file the moment your assistant writes it and hands back only the error-level alerts, so mistakes get fixed in the same turn they were made.
See the hookThe engine itself, as tools an assistant can call — so a rule it writes gets compiled, tested, and costed before it reaches your repository.
What it can checkOther clients can read skills/ directly.
/plugin marketplace add vale-cli/agent-tools
/plugin install vale@agent-toolsThe skills and the hook need nothing but the Vale binary. The MCP server is the one paid piece — it runs on Vale CMS, and everything else works without it.