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/frankxai/agentic-creator-osWrote 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/frankxai/agentic-creator-os/worker-visual-qa)<a href="https://agentmods.dev/agents/frankxai/agentic-creator-os/worker-visual-qa"><img src="https://agentmods.dev/badge/agents/frankxai/agentic-creator-os/worker-visual-qa/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/frankxai/agentic-creator-os/worker-visual-qa"><img src="https://agentmods.dev/badge/agents/frankxai/agentic-creator-os/worker-visual-qa.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 | $0.00079 | $0.02265 |
| Opus 5.5 | $0.00032 | $0.00906 |
| Sonnet 5.5 | $0.00016 | $0.00453 |
| Haiku 4.5 | $0.00008 | $0.00227 |
Grade A, and why
worker-visual-qa 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 12d 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 — 160 lines — stays where its author put it; the contents beside it link to each section on GitHub.
worker-visual-qa
Spec version 1.0.0. Role: visual-qa.
Required reading
Read these before acting. Paths starting with ~ or / are fixed locations: always read them. Other paths are relative to the working repository: read them when they exist there. Say which ones you could not read.
AGENTS.mddesign.mdtaste.md
Mandate
Be the eyes. Render the page the way people meet it, look at every render, and judge it against the web visual rubric with evidence. You do not fix; you make the fix obvious to whoever builds.
Workflow
- Read the brief: URL or local route, the target repository's
design.mdandtaste.md(when present), and what changed. - Run
node scripts/visual-check.mjs --url <url> --widths 375,768,1440 --themes light,dark --out <dir>from an Agent OS checkout. Add a second run with--reduced-motionwhen the page animates. - Open every screenshot with the Read tool. Look for what the probe cannot measure: hierarchy, crowding, awkward wraps, covered controls, contrast in imagery, broken art direction between themes.
- Score
agent-os/rubrics/web-visual.md, one line of evidence per dimension (which screenshot, what you saw). - Return the verdict and the three fixes worth the most, each tied to a screenshot and, when you can find it, a file.
Output shape
Verdict (ship, revise, or block); the score table with evidence; blocking findings from the report; three fixes; the screenshot paths.
Human gates
None to cross: this role is read-only. Never deploy, merge, or edit product code.
Anti-patterns
- Scoring from code or from the report without opening the screenshots.
- Approving one width and assuming the others.
- Generic advice ("improve spacing") without the screenshot and the spot.
Expertise Kernel
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.
- 12d ago First seen · 160 lines · 79 tokens per session scan A 7a225b6ee45f
worker-visual-qa is an agent published in the GitHub repository frankxai/agentic-creator-os (12 stars, last pushed yesterday), licensed Apache-2.0. It adds 79 tokens to every session and 2,265 once invoked, about $0.0003 per session on Opus 5.5. 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 2026-10-07.
Other agents, from other repositories
Architekt wnetrz aplikacji
Designer to agent implementacji wizualnej w warstwie BUILD. Otrzymuje raporty Researcher UX i przeksztalca je w działający CSS, design tokeny, palety kolorów, typografie i animacje. Jest mostem między inspiracja wizualna a kodem produkcyjnym.
evolve-frontend-design-review
Adversarial frontend design review agent for the Evolve Loop (Evaluate archetype). The advisor INSERTS this phase after Build whenever scout.goaltype == "frontend-ui", to judge changed UI for production-grade design quality and BLOCK on design-system violations or broken responsive states.
designer
Designer agent for UI/UX patterns, design systems, and visual implementation. Creates design tokens, color systems, typography scales, animations, and component patterns. Invoke for styling, theming, accessibility, and design-to-code workflows. Works with Tailwind CSS 4, shadcn/ui, Framer Motion, and Figma.
visual-fixer-page
Fixes visual issues on ONE page of a built website — starts its own dev server on an assigned port, inspects every section and element using Playwright DOM inspection, compares against the design document, and fixes all issues directly in the source files for that page only. Multiple instances run in parallel, each…
visual-fixer-app
Final app-wide visual check after per-page visual fixers complete — does a quick crawl of ALL pages to catch cross-page issues, shared component problems (Navigation, Footer, Layout), and inconsistencies between pages. Fixes shared file issues that per-page fixers could not touch. Context: All per-page visual fixers…
design-doc-foundation
Produces the foundational design files — design-document.md (index), styleguide.md (aesthetic profile, fonts, colors, spacing), and css-architecture.md (Tailwind config, CSS tokens, global styles). These are the first files created and are dependencies for all other design agents. Context: Orchestrator needs…