From rolepod-uiproof
Measure Core Web Vitals (LCP, INP, CLS) on a live page using PerformanceObserver injection. Returns metrics + thresholds verdict per web.dev good/needs-improvement/poor bands. Chromium-only — Firefox and WebKit ship partial PerformanceObserver coverage.
How this skill is triggered — by the user, by Claude, or both
Slash command
/rolepod-uiproof:measure-cwvThe summary Claude sees in its skill listing — used to decide when to auto-load this skill
Single-backend skill. Calls **`measure_cwv`** on the rolepod-uiproof
Single-backend skill. Calls measure_cwv on the rolepod-uiproof
MCP server. No fallback (D-024).
/audit-a11y and /check-errors.performance-engineer agent territory.cwv_unsupported_browser.url — URL to measure.viewport — optional { width, height }.emulate.network_throttle — one of offline | slow-3g | fast-3g | slow-4g | fast-4g | no-throttling.emulate.cpu_throttle — 1 (none) to 20 (very slow). Chromium only.observe_ms — observation window after load. Default 5000ms.interactions[] — optional steps to elicit INP samples. Each step is
{ kind: "click", query }, { kind: "type", query, text },
{ kind: "key", key }, or { kind: "scroll", direction, amount? }.thresholds — overrides for lcp_ms (2500), inp_ms (200), cls (0.1).run_id — folder under the run root (see Evidence routing).metrics — { lcp_ms, inp_ms, cls }.verdict — per-metric good | needs-improvement | poor | unmeasured. INP
is unmeasured when no interactions were driven.status — pass (all good) | warn (any needs-improvement) | fail (any poor).thresholds — the effective thresholds used.report_path — path to cwv.json with full samples.measure_cwv input from the user's intent (URL, optional
throttle, optional interactions).Run artifacts are saved under:
.rolepod-uiproof/artifacts/measure_cwv_<ts>_<uuid>/rolepod parent (detected via the marker file <git-root>/.rolepod/parent-active written by the parent's SessionStart hook): <git-root>/.rolepod/evidence/<ts>-rolepod-uiproof-measure-cwv/Either way the run directory contains a manifest.json per Extension Protocol v1 so the parent's check-work skill can aggregate the result into the verify report. Standalone users can read the manifest themselves — same shape.
Surface plainly:
The
/measure-cwvskill needs the rolepod-uiproof MCP server, which is not currently available. Confirm the plugin is installed and try again.
Do not attempt another backend (D-024).
{
"url": "https://example.com",
"observe_ms": 5000
}
INP will be reported as unmeasured — no interactions were driven.
{
"url": "https://example.com/app",
"observe_ms": 3000,
"interactions": [
{ "kind": "click", "query": "Open menu" }
]
}
{
"url": "https://example.com",
"emulate": { "network_throttle": "slow-4g", "cpu_throttle": 4 },
"observe_ms": 7000
}
Guides creation, editing, and verification of skills for AI coding agents using test-driven development with subagent scenarios. Use when authoring or debugging skills.
npx claudepluginhub nuttaruj/rolepod-uiproof --plugin rolepod-uiproof