Storefront → Templates → Migration validation
Commands reference
Every parity command — inspect probes, run management, and the full flag surface.
The full command list. The flagship commands have their own pages (run, migrate, fix/section, vitals/cache, e2e/audit, journey, extract, pr); the rest are here.
Inspect — sub-10s single-purpose probes
| Command | What it does |
|---|---|
css-trace --selector <s> [--url | --prod/--cand] | List every CSS rule affecting an element, or diff computed styles prod vs cand. --filter, --json. |
console --url <u> | Capture console errors/warnings + network failures on one page. --filter error,warning, --wait, --json. |
html --url <u> [--prod/--cand --diff] --selector <s> | Dump a page/selector's HTML, or a unified prod × cand diff. --pretty, --json. |
check <name> --prod --cand | Run ONE named check on a single page (skips sitemap + LLM). See Checks. |
Manage — saved runs & artifacts
| Command | What it does |
|---|---|
list / list modules --json | List saved runs; list the 8 selectable modules. |
report <runId> [--section <name>] [--json] | Reopen a run's HTML report, or extract one tab (HTML / JSON projection). |
serve <runId> [--port] | Local server for a run's report + an iframe proxy that strips X-Frame-Options/CSP for the side-by-side view. |
compare <runId> --against <baseline> | Compare a run to a saved baseline. |
baseline set <runId> --name / list / unset <name> | Manage baselines. |
prompt <runId> [--min-severity] [--limit] | Export a run's ranked issues as a Markdown prompt for any LLM. |
explain <runId> <issueId> | LLM deep-dive on one issue (needs an API key). |
learned stats / learned validate --url | Inspect / validate the learned-selectors library. |
Run parity <command> --help for the exact, current flags — this page summarizes; --help is authoritative.