Getting it into your agent
One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.
npx skills add n-n-code/n-n-code-skills --skill lighthousegit clone --depth 1 https://github.com/n-n-code/n-n-code-skillsWrote this? Show the measurements
A badge with what this costs and how it scanned, read live from this page, so it follows the numbers instead of freezing them. Markdown for a README, HTML for a documentation site or a project page.
[](https://agentmods.dev/skills/n-n-code/n-n-code-skills/lighthouse)<a href="https://agentmods.dev/skills/n-n-code/n-n-code-skills/lighthouse"><img src="https://agentmods.dev/badge/skills/n-n-code/n-n-code-skills/lighthouse.svg" alt="Measured on agentmods" height="20"></a>What it costs to keep this loaded
Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.
| Model | Per session | Once invoked |
|---|---|---|
| Fable 5.1 | $0.00071 | $0.01452 |
| Opus 5 | $0.00036 | $0.00726 |
| Sonnet 5 | $0.00014 | $0.00290 |
| Haiku 4.5 | $0.00007 | $0.00145 |
Grade A, and why
lighthouse scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured today.
A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.
Nothing flagged
None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.
How it starts
The opening of the file, as written. The whole thing — 108 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Google Lighthouse
Turn a page-quality question into reproducible measurements, actionable findings, and verification appropriate to the claim. This is a portable workflow skill. It uses upstream Lighthouse and Lighthouse CI (LHCI), without a bundled runner or required host-specific interface. Reading an existing report needs no browser; native execution needs a compatible Node runtime, Lighthouse, and Chrome/Chromium. Scripted flows additionally need Puppeteer; CI collection needs LHCI.
Choose the job and owner
| Requested job | Work and output | Change boundary |
|---|---|---|
| Interpret existing reports | Explain findings, compare compatible evidence, identify gaps | Read supplied artifacts; no new audit or source changes unless requested |
| Audit a URL or page state | Execute the relevant measurement and retain local reports | Browser activity and report files within the requested scope |
| Improve measured behavior | Inspect causes, implement requested fixes, rebuild and remeasure | Scoped source changes under existing authorization |
| Configure CI | Establish collection, assertions, and artifact handling | Requested tooling, configuration, and pipeline changes |
Let this skill own Lighthouse measurement and interpretation. Use
chrome-devtools-axi for general Chrome investigation and AXI execution rules;
honor an explicitly chosen browser, wrapper, or direct tool interface. Use
playwright-testing for existing Playwright tests and setup-playwright for
their harness. Add matching implementation guidance or ui-guidance /
ui-design-guidance when fixing code or reviewing broader UI/accessibility
behavior. Add tester-mindset only when the validation strategy needs framing.
These companions are optional; Lighthouse use alone does not require a harness.
Establish, measure, interpret, verify
- Establish the claim and inputs. Identify the requested job, URLs or report files, page states, device scope, authentication, and any existing baseline or budgets. Inspect repository scripts, lockfiles, Lighthouse/LHCI configuration, and CI before asking about facts already recorded there.
- Select a supported surface and mode. Prefer the CLI for reproducible URL audits and the Node API for scripted flows. Check versions, prerequisites, runtime help, and supported categories. A wrapper may expose only part of Lighthouse. Choose navigation for a page load, timespan for a bounded interaction, or snapshot for the current DOM state. Read execution and configuration before running; read authentication and flows when state or interactions matter.
- Define comparable conditions. For an unspecified URL audit, default to mobile navigation and the supported standard categories: performance, accessibility, best practices, and SEO. Prefer production assets and the repository's real startup contract. Record build/commit, Lighthouse and browser versions, mode, viewport/form factor, throttling, cache/storage, auth state, and material environment differences. Label development-build evidence when that is the relevant or only available target.
- Run and validate the capture. Use a fresh output prefix per attempt; save JSON and HTML where supported. Verify process status, report identity, time, actual destination/page state, runtime errors, and warnings before interpreting scores. A successful process or an existing file is insufficient. For performance comparisons, collect three sequential runs per variant on the same apparatus; keep failed attempts visible and report valid sample counts, medians, and variation. Increase sampling only when noise can change the conclusion. Do not run competing audits on the same machine.
- Interpret before proposing changes. Use reports and improvements. Separate measurements, observed resource/element evidence, and causal hypotheses. Distinguish absent, null, manual, informational, not-applicable, and error results. Check audit IDs against the report's version; missing legacy audits are not proof of a fix. Prioritize user impact and demonstrated causes.
- Act within the selected job. For improvement work, make the scoped fix, rebuild/restart owned services as needed, and repeat the same measurement plus relevant functional or UI checks. Stop at sufficient evidence for the requested outcome; explain noise, remaining issues, and blocked targets. For CI work, use Lighthouse CI; preserve established budgets and calibrate new hard gates from a baseline.
- Report and finish ownership. Provide the outcome, conditions, valid and failed runs, category scores and metric units, prioritized findings, verified changes, artifact paths, and remaining limits. Record the exact invocation or configuration needed to reproduce the result. Stop only owned processes and restore task-changed conditions in a reused session.
What ships with it
5 files beside SKILL.md in the same directory: the scripts, references and assets a skill reads on demand. Not counted in the per-session cost; read them before you install if any of them is executable.
What this file has done since we first saw it
Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.
- today First seen · 108 lines · 71 tokens per session scan A d2bde8c07c38
lighthouse is a skill published in the GitHub repository n-n-code/n-n-code-skills (4 stars, last pushed yesterday), licensed MIT. It adds 71 tokens to every session and 1,452 once invoked, about $0.0004 per session on Opus 5. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-09-06.
Other skills, from other repositories
web-frontend-designer
Scaffold frontend architecture, components, routing, and UI dependencies from project config and wireframes.
web-frontend-tester
Run frontend functional, accessibility, visual, and performance test planning with structured reports.
baoyu-markdown-to-html
Converts Markdown to styled HTML with WeChat-compatible themes. Supports code highlighting, math, Mermaid (rendered to PNG via headless Chrome), PlantUML, footnotes, alerts, infographics, and optional bottom citations for external links. Use when user asks for "markdown to html", "convert md to html", "md 转 html"…
frontend-specialist
A front-end development workflow for building or improving web pages, components, layouts, forms, dashboards, and responsive interfaces. It covers common tools and frameworks such as React, Vue, Next.js, Tailwind, and CSS.
migrate-divi-to-breakdance
Use when the user says 'migrate divi to breakdance', 'convert divi to breakdance', or 'rebuild divi pages in breakdance'. Parses Divi shortcodes, maps modules to Breakdance elements, and creates draft duplicates for review.
migrate-divi-to-bricks
Use when the user says 'migrate divi to bricks', 'convert divi to bricks', or 'rebuild divi pages in bricks'. Parses Divi shortcode content, maps modules to Bricks elements, and generates draft duplicates for review.