manni meta

manni meta checks that your documents carry the metadata your tooling depends on. It checks that every page declares a type, that timestamps are real dates, and that required fields are actually present. It fails your build when they aren’t. It reads the frontmatter or headers from your files, validates them against a JSON Schema, and reports what’s missing or malformed.

It validates the presence and format of metadata. It does not grade your prose, check your spelling, or judge your content. If your schema says every page needs a title and an ISO 8601 timestamp, manni meta enforces exactly that and nothing more.

A 30-second proof

You don’t need to install anything to try it. Point manni meta at a Markdown file:

npx @hawkeyexl/manni meta validate path/to/your-doc.md

A file with complete, well-formed metadata passes:

✓ path/to/your-doc.md

1 file checked, 1 passed, 0 failed, 0 errors

A file that’s missing a required field fails, and tells you which field, on which line, and which schema flagged it:

✗ path/to/your-doc.md
    (root)  must have required property 'type'  [google:okf:0.1]

1 file checked, 0 passed, 1 failed, 1 error

That’s the whole loop: run it, read the result, fix the field, run it again. With no schema configured, manni meta validates against its built-in OKF schema, which requires only type.

What do you want to do?

Set up validationAdd a config file and a schema, then wire manni meta into your repo. Start here if you own the docs.

Run it in CIDrop manni meta into GitHub Actions, GitLab CI, Jenkins, or pre-commit. Exit codes and PR annotations.

Define & evolve schemasEncode your metadata standard as a JSON Schema, wire it to the right files, and evolve it safely.

Fix a failing checkYour PR's manni meta check is red? Map the error to a field and line, fix it, and confirm green.

ReferenceEvery command, flag, config key, output format, and exit code. The lookup shelf.

Not sure you need to write a schema at all? manni meta ships twenty-three. They run from the Hugo, Jekyll, Starlight, Docusaurus, Antora and Sphinx front matter contracts to the Diátaxis and Good Docs vocabularies. They also cover Dublin Core, Open Graph, and the Agent Skills SKILL.md format. The built-in schema registry lists them all in one table.

What manni meta checks, and what it doesn’t

It checks

Required fields are present. Values match the format your schema declares (uri, date-time, enums, types). Metadata stays consistent as your repo grows.

It doesn't

Grade writing quality, check grammar or links, render your docs, or care about anything below the frontmatter. Metadata in, pass/fail out.

New here? The Get started page installs manni meta and walks you through your first passing and failing result in under a minute.