Lima for VS Code

Syntax highlighting and live diagnostics for Lima — the small, specified frontmatter format.

What it does

Highlighting

.lima files — a full TextMate grammar: keys, comments, block (-) and flow ([ ], { }) collections, | literal block scalars and ^^ continuation lines, scalar types (string, number, boolean, null, date in ISO 8601 / DD.MM.YYYY / DD/MM/YYYY form), and single/double-quoted strings with their real escape rules.
References 2.0 tokens — ${document.path} and $(partial.path) are highlighted where they are active (unquoted scalars, flow values); they stay literal inside quoted strings, matching the spec.
Markdown / MDX frontmatter — the --- … --- block at the top of a .md or .mdx file is highlighted as Lima.

Diagnostics

The real Lima parser runs as you type — over .lima files and the frontmatter block of Markdown / MDX — and reports the specification's own diagnostics: duplicate keys, invalid dates, unknown escape sequences, unterminated strings, bad flow syntax, resource-limit violations, and the rest, each with its diagnostic code.

Checking is strict by default (an editor is where you want that feedback). Settings:

Setting	Default
`lima.diagnostics.enable`	`true`	Turn diagnostics on/off.
`lima.diagnostics.strict`	`true`	Strict mode. Off = match a non-strict runtime, which coerces instead of erroring.
`lima.diagnostics.ignoreUnresolvedReferences`	`true`	Suppress unresolved-reference / partial findings — the editor has no cross-file or `partials` context.

Not part of Lima (and so not highlighted)

> folded block scalars — use | plus ^^.
YAML anchors, aliases, tags, merge keys, directives, multi-document streams.
References 1.0 ($key) / (%key) syntax — in References 2.0 these are ordinary string content.

Development

bun install         # also builds dist/ (via the `prepare` script)
bun test            # grammar scopes + diagnostic findings
bunx tsc --noEmit   # typecheck src/
bun run build       # rebuild dist/extension.js
bun run watch       # rebuild on change

dist/ is a build artifact and is git-ignored. bun install regenerates it through the prepare script, so a fresh checkout (or a git clean) plus bun install leaves a working extension. bun run watch while developing.

The parser is bundled from ../../js/src so the checker and the published @limaformat/lima package never drift. Grammar lives in syntaxes/.

License

ISC — see LICENSE. Part of the limaformat/lima monorepo (editors/vscode).