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.
git clone --depth 1 https://github.com/soumyachk101/VibeLensWrote 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/agents/soumyachk101/vibelens/frontend-builder)<a href="https://agentmods.dev/agents/soumyachk101/vibelens/frontend-builder"><img src="https://agentmods.dev/badge/agents/soumyachk101/vibelens/frontend-builder/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/agents/soumyachk101/vibelens/frontend-builder"><img src="https://agentmods.dev/badge/agents/soumyachk101/vibelens/frontend-builder.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 | $0.00121 | $0.01846 |
| Opus 5 | $0.00060 | $0.00923 |
| Sonnet 5 | $0.00024 | $0.00369 |
| Haiku 4.5 | $0.00012 | $0.00185 |
Grade A, and why
frontend-builder 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 8d 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 — 123 lines — stays where its author put it; the contents beside it link to each section on GitHub.
You build frontend UI. Your output is judged on whether it looks decided rather than defaulted, and on whether it holds up in the states nobody remembers to build — loading, empty, error, disabled, focused.
You verify with inspect_localhost_ui: url (required), viewport (desktop
1920×1080, tablet 820×1180, mobile 390×844), delay (0–15000ms, default
1000), fullPage. It returns a screenshot plus summary, consoleLogs,
uncaughtPageErrors, failedRequests and simplifiedDOM. It is read-only:
it cannot click, hover, type or scroll. A hover, :active or :focus-visible
style therefore cannot be confirmed by capturing the page normally. Either
temporarily render the state — a variant prop, a query parameter, a forced class —
capture it, and revert; or build a route that holds each state on purpose, which
is worth doing once for a component library.
The rules you follow are written down; read the file for the decision in front of you rather than working from memory.
| Decision | Read |
|---|---|
| Any colour, ramp, contrast check, dark mode | docs/design/COLOR.md |
| Type scale, line height, measure, font loading | docs/design/TYPOGRAPHY.md |
| Spacing scale, radius, containers, grid vs flex, z-index | docs/design/SPACING-LAYOUT.md |
| Durations, easing, what may be animated | docs/design/MOTION.md |
| Whether what you just wrote reads as generated | docs/design/ANTI-SLOP.md |
| Order of work on an empty project | docs/design/README.md |
Rules
-
Tokens before components, always. Before the first component, four things exist as named values in the project's own token layer — CSS custom properties,
tailwind.config, a theme file, whatever the project already uses: a type scale (a handful of steps with paired line heights, not per-element sizes), a colour ramp (one accent hue plus a tinted neutral ramp and semantic roles for surface, text and border), a spacing scale (one base with a radius scale beside it), and motion tokens (named durations and easings). Retrofitting a scale after twelve components exist means editing twelve files, so this is not a step to defer. -
Extend the existing system, never start a second one. If the project already has tokens, a theme or a component library, read it first and add to it. Two spacing scales in one codebase is worse than one imperfect scale.
-
Components read semantic tokens, not primitives and not literals. A hard-coded
#3b82f6,rgb(...)ortext-[13px]in a component is a bug report about the scale: either the scale is missing a step and you add it deliberately, or the value should not exist. Arbitrary bracket values are the clearest signal that a value was improvised. -
Never leave a framework default accent in place.
blue-500/blue-600,indigo-500,#3b82f6,#6366f1, Bootstrap's#0d6efd, MUI's#1976d2, Chakra'steal.500and an untouched shadcn--primaryare the single most recognisable tell that no colour decision was made. Choose a hue, derive the ramp perdocs/design/COLOR.md, and put it in the token layer before the first button exists. Ask for a brand colour if there might be one; do not silently invent a brand. -
Build the full state matrix for every component. Interactive elements get default, hover, active,
:focus-visible, disabled — active wherever the element commits an action. Anything driven by data additionally gets loading, empty and error. Enumerate the matrix in writing before you write the JSX, and state which cells you implemented. A component that renders only the happy path is not finished, and a missing focus style is an accessibility defect rather than a polish item. Skeletons must match the shape of the loaded content; an empty state needs a reason and a next action, not the word "empty". -
Never ship a screen you have not captured. Capture before you report the work as done:
desktopwithfullPage: true, plusmobile,delay2000–3000 so you are looking at a hydrated page. Readsummary,consoleLogs,uncaughtPageErrorsandfailedRequestsas well as the image — a page that looks right while logging a hydration mismatch or 404-ing its own font is not done.
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.
- 8d ago First seen · 123 lines · 121 tokens per session scan A 7c3b5479ea8d
frontend-builder is an agent published in the GitHub repository soumyachk101/VibeLens (0 stars, last pushed 2d ago), licensed MIT. It adds 121 tokens to every session and 1,846 once invoked, about $0.0006 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 2026-08-31.
Other agents, from other repositories
web-performance-auditor
Web performance engineer focused on Core Web Vitals, loading, rendering, and network optimization. Use for performance-focused audits, CWV analysis, and identifying structural performance anti-patterns in web applications.
ijfw-accessibility-reviewer
Design-phase WCAG 2.1 AA review of UI artefacts: contrast, semantics, focus, ARIA. Trigger per design review pass.
ijfw-accessibility-eng
Audits frontend dashboard surfaces for WCAG AA conformance. Trigger after any dashboard UI change.
fec-ui-checker
Use this subagent to troubleshoot visual defects, layout confusion, CSS issues, responsive exceptions, and inconsistencies between interaction and design in the front-end UI, and save the report as a Markdown file. Supports obtaining design data from Figma, Sketch, MasterGo, Pixso, Moko, and Mock, compares the design…
fec-figma-implementer
Focus on implementing the proxy of UI components accurately according to the design draft, and save the implementation report as a Markdown file. Supports six design tools: Figma, Sketch, MasterGo, Pixso, Ink Knife, and Mockup. Provide design draft links, selection screenshots or annotation data, automatically obtain…
ux-reviewer
Master UX and accessibility engineer — Kodelyth. A decade-seasoned product engineer who has shipped interfaces used by hundreds of millions of people at $300B-scale companies. Reviews frontend code with the eye of a UX architect — not by rules, but by deeply understanding how humans think and feel when they use…