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 aziontech/webkit --skill webkit-ui-verifygit clone --depth 1 https://github.com/aziontech/webkitWrote 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/aziontech/webkit/webkit-ui-verify)<a href="https://agentmods.dev/skills/aziontech/webkit/webkit-ui-verify"><img src="https://agentmods.dev/badge/skills/aziontech/webkit/webkit-ui-verify/github.svg" alt="Measured on agentmods" height="20"></a>Or the 80×15 button, for a site that already has a row of RSS and ATOM ones. Only the verdict fits; the numbers stay here.
<a href="https://agentmods.dev/skills/aziontech/webkit/webkit-ui-verify"><img src="https://agentmods.dev/badge/skills/aziontech/webkit/webkit-ui-verify.svg" alt="Reviewed on agentmods" width="80" 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 | $0.00000 | $0.02472 |
| Opus 5.5 | $0.00000 | $0.00989 |
| Sonnet 5.5 | $0.00000 | $0.00494 |
| Haiku 4.5 | $0.00000 | $0.00247 |
Grade A, and why
webkit-ui-verify 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 1mo ago.
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 — 184 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Skill: webkit-ui-verify
Purpose
The other skills in this pack prescribe — tokens for both themes, a full state surface, WCAG at the composition layer, layout across breakpoints. This one verifies they held, by driving the running screen in a real browser and looking. A screen "looks done" only in the one theme, one width, and one data state you happened to open; the regressions the other skills warn about — a dark-mode-only contrast break, an empty fetch that renders a blank panel, a focus that never returns to its trigger, a Vue warning in the console — are invisible until something exercises them. This is the executable companion to the whole pack: the runtime gate you run before calling any route finished.
The method is reconnaissance, then action — OBSERVE, never assume. Navigate, screenshot, read the live accessibility tree, and let what is actually on the page drive the next step. Do not assert against selectors or state you expect to exist; find them, then act.
How to use
/webkit-ui-verify— run the full pass against the route you just built in this conversation./webkit-ui-verify <route-or-file>— verify a specific route (/workloads) or the view a.vuefile renders; report per finding: what failed, the evidence (console line, axe rule, screenshot), and the fix.
The agent form is webkit-ui-verifier — hand it a route and it runs this loop end to end and reports back. Either way the loop is the same; only who drives it changes.
Before you run
- Build/serve the app locally first (
pnpm dev/pnpm build && pnpm preview), and verify the URL responds — the pass drives a live server, not a static file. - The consumer installs two devDeps:
playwright(the browser) andaxe-core(the a11y engine). Install Chromium once withnpx playwright install chromium.
The loop — capture, then assert
Run every route through all four, in both themes and at 2–3 widths. A pass that skipped a theme, a width, or a state did not verify the screen.
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.
- 1mo ago First seen · 184 lines · 0 tokens per session scan A 14b901c76b93
webkit-ui-verify is a skill published in the GitHub repository aziontech/webkit (3 stars, last pushed today), licensed MIT. It costs nothing until one of its globs matches a file; then it loads 2,472 tokens. 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 2026-08-31.
Other skills, from other repositories
writing-browser-e2e-tests
Playwright E2E testing standards from authoritative docs/explanation/software-engineering/automation-testing/tools/playwright/ documentation.
sigil-audit
The auditor crawls every /docs/components/ page on the local dev server and combines static MDX analysis with a real Chromium browser pass.
exploratory-testing
Guides session-based exploratory testing of a running interface or API: framing charters, choosing tours, citing a specification, contract, or design for every expected result, telling a defect from a gap, and staying non-destructive.
web-testing-exploratory-methodology
Spec-aware session-based exploratory testing methodology for swe-web-tester (exploratory charter) — charter framing, testing tours, SFDIPOT/CRUSSPIC STMPL coverage models, the test dimensions checklist, three mandatory systematic sweeps, specs-as-ground-truth comparison and spec-gap detection, defect anatomy, and the…
browse
Drive a real browser through Aside: open a page, read it, click through a flow, take screenshots, check console errors. (gstack).
cua-driver
Use Cua Driver for desktop or browser tasks that are awkward or unavailable through Bash/APIs, or when the user explicitly wants GUI interaction: app testing, visual bug reproduction, form filling, calendar entry, screenshots, and demo recording. Also covers Cua setup; not OpenAI Codex Computer Use or web research.