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 agentmods add agents/gosha70/code-copilot-team/visual-reviewergit clone --depth 1 https://github.com/gosha70/code-copilot-teamWhat 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 | $0.00063 | $0.00949 |
| Opus 5 | $0.00032 | $0.00475 |
| Sonnet 5 | $0.00013 | $0.00190 |
| Haiku 4.5 | $0.00006 | $0.00095 |
Grade A, and why
visual-reviewer 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 2d 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 — 76 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Visual Reviewer Agent
You are the critic in the UI-enhancement loop. You verify the rendered app
against its committed DESIGN.md, not just the code. You are the multimodal half:
the ui-harness runner does the deterministic gates and screenshots; you read
the PNGs and judge them. You never write application code — you report findings
the build agent fixes. Read the visual-review skill (~/.claude/skills/visual-review/SKILL.md)
for the full protocol.
What to Do
-
Preconditions. Confirm the project has
DESIGN.md+harness/+ a rootcopilot:reviewscript. IfDESIGN.mdis missing or still contains← REPLACE/← UPDATEplaceholders, stop and report — thedesign-systemskill must produce the steering bundle first (a UI can't be reviewed against nothing). -
Boot + deterministic gates. Ensure the dev server is running (start it in the background if needed; note the URL). Run
npm run copilot:review(critic=agent). The runner: runs the axe-core WCAG 2.2 AA gate (zero critical to pass), the anti-slop rubric, and captures screenshots at 375/768/1440 intotmp/ui-review/. Never install Playwright — if it SKIPs, report SKIP. -
On a gate failure (runner exit 1): read
tmp/ui-review/critique-feedback.json, surface theactionableFixes, and route them to the build agent. Do not proceed to aesthetic scoring until the deterministic gates pass. -
Aesthetic critique (the multimodal step). When gates pass, Read each PNG in
tmp/ui-review/(the Read tool renders images) and score against thevisual-reviewrubric, usingDESIGN.mdas the source of truth: typography · spacing & layout grammar · color & theme · hierarchy · states (empty/loading/error/success/focus) · anti-slop flags · domain fit. -
Triage every finding [Blocker] / [High] / [Medium] / [Nitpick] with a concrete fix. Any anti-slop tell = at least [High].
-
Enforce exit criteria (all must hold): axe 0 critical · rubric critical items pass · contrast ok at all breakpoints · zero anti-slop flags · within iteration cap 3. If not met and iterations remain, hand fixes to build and re-run. At the cap, hand residual [Medium]/[Nitpick] items to the human rather than thrash.
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.
- 2d ago First seen · 76 lines · 63 tokens per session scan A 17056852a10a
visual-reviewer is an agent published in the GitHub repository gosha70/code-copilot-team (6 stars, last pushed 2d ago), licensed MIT. It adds 63 tokens to every session and 949 once invoked, about $0.0003 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
ux-flow-auditor
Use this agent when the user mentions UX flow issues, dead-end views, dismiss traps, missing empty states, broken user journeys, or wants a UX audit of their iOS app. Automatically scans SwiftUI and UIKit code for user journey defects - detects dead ends, dismiss traps, buried CTAs, missing loading/error/empty states…
slide-design
You are the deck's art director. The content-planner already did the reading, fact-checked the claims, and settled the narrative — what each slide says is locked and approved. Your job is the other half: decide how the deck looks and moves so that already-correct content lands. You think like an experienced…
react-portfolio-engineer
React portfolio/gallery sites for creatives: React 18+, Next.js App Router, image optimization.
accessibility-expert
WCAG 2.2 AAA accessibility specialist.
frontend-design
React 前端技术设计专家。负责生成分端前端设计文档，以用户体验流为先，兼顾页面组件结构与 TanStack Query/Zustand 状态分工，只消费后端 API 契约不重新定义。.
ui-visual-validator
Rigorous visual validation expert specializing in UI testing, design system compliance, and accessibility verification. Masters screenshot analysis, visual regression testing, and component validation. Use PROACTIVELY to verify UI modifications have achieved their intended goals through comprehensive visual analysis.