RMT Linter

Diagnostics, JSON output and agent repair reports for RMT sources.

What it covers

The RMT linter reads source, parser diagnostics, and semantic rules, then returns stable machine-readable findings. It does not modify source files or compile an apparently valid app after a severe error.

Public building blocks

• tools/rmt-linter/cli.js is the command-line entry point.
• tools/rmt-language/diagnostics.js normalizes codes, fields, and severity.
• tools/rmt-language/rules/ contains rule-based checks.

• xt rmt lint app.rmt.
• --json for tools.
• --agent for repair reports.

CLI

xt rmt lint app.rmt
xt rmt lint app.rmt --json
xt rmt lint app.rmt --agent
xt rmt lint app.rmt --fail-on warning
node scripts/run_xtend_tests.js rmt-language-regression --json

The agent report contains repairPlan, fixOrder, relatedDiagnostics, explained No-Op entries and diagnostics such as rmt.document.extension.fallback-used. The public schema is xtend.rmt.tooling-docs.v1; Reports can use xtend.rmt.linter.report.v1 and xtend.rmt.ai-agent-repair-report.v1.

AnimationEngine diagnostics cover unknown effects, invalid interrupt or reducedMotion policies, missing layoutKey for shared-element/layout-flip, unsafe keyframe properties and missing filter opt-in for fade-blur. Code actions can replace unsafe enum values with safe defaults and add missing motion policy fields when the diagnostic points at a concrete record.

Recommended workflow

Lint source before compiling. Resolve syntax errors first, then unknown references and policies; apply an automatic fix only when the report identifies the exact text edit.

Next steps

• XTendRMT overview
• RMT Authoring Guide
• RMT Reference
• RMT Language Server