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 mtarcure/claude-vibe-squad --skill site-preship-reviewgit clone --depth 1 https://github.com/mtarcure/claude-vibe-squadWrote 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/mtarcure/claude-vibe-squad/site-preship-review)<a href="https://agentmods.dev/skills/mtarcure/claude-vibe-squad/site-preship-review"><img src="https://agentmods.dev/badge/skills/mtarcure/claude-vibe-squad/site-preship-review/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/mtarcure/claude-vibe-squad/site-preship-review"><img src="https://agentmods.dev/badge/skills/mtarcure/claude-vibe-squad/site-preship-review.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 | $0.00062 | $0.01144 |
| Opus 5.5 | $0.00025 | $0.00458 |
| Sonnet 5.5 | $0.00012 | $0.00229 |
| Haiku 4.5 | $0.00006 | $0.00114 |
Grade A, and why
site-preship-review scanned grade A with 1 finding against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured yesterday.
A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.
Makes network callslowCapability
Not a fault in itself. Listed so you know the mod talks to something, and to what.
`NEXT_PUBLIC_`/`VITE_` leaks); security headers via `curl -sI -L` (HSTS, nosniff, frame-options, a CSP that works); How it starts
The opening of the file, as written. The whole thing — 63 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Site pre-ship review
Six reviewers, one dimension each, run in parallel against a frozen branch of the site. Each renders every page and LOOKS at it; none edits. Then a single fix worker applies every BLOCKER and MAJOR plus the quick MINORs, and a second fix pass takes whatever reviews landed late. Idea credited to an @voxyz_ai post (five subagents, twenty checks); this is our own checklist, tuned by what actually broke.
Packet shape (shared by all six)
- Point at a branch, never a working tree:
git worktree add --detach /private/tmp/site-review-$RANDOM <branch>and serve withpython3 -m http.server. - Render with Python Playwright launching its own headless Chromium at 1440x900 and 390x844 (plus 360/320 for headers). Never attach to the operator's browser.
- Give the reviewer the product facts the copy must match (prices, limits, chains, tagline) and a read-only pointer to the code that is the source of truth for them.
- Deliverable: numbered findings for that dimension only (page + element + screenshot path + BLOCKER/MAJOR/MINOR + exact fix), then PASS/FAIL per checklist item. Screenshots go OUTSIDE the repo; a worker that commits them to the site repo gets its whole commit refused at integration.
The six checklists
- Design — one type scale and casing; consistent spacing grid; brand identical on every page; WCAG AA contrast and an obvious primary call to action on every page, mobile included.
- Mobile (390 and 360) — no horizontal scroll (long addresses/code wrap inside their box); tap targets >= 44 px; menu opens, closes, every item works; body text >= 16 px, nothing clipped.
- States — every form's empty/invalid/success/failure states, with the network request intercepted (never send a real message); 404; copy-button feedback, hover/focus/disabled, visible keyboard focus; no console errors; works with JS disabled or slow.
- User flows — the five-second test, then Get Started → install → first action, checked against the real code; pricing → start; every internal link and anchor resolves; legal and support reachable from every footer.
- Launch prep — unique title, description, canonical, og:* (1200x630 image that exists), twitter:card on every page; favicon, robots.txt, sitemap.xml; no placeholder/TODO/staging/testnet copy or internal paths/IDs; page weight and render-blocking assets.
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.
- yesterday First seen · 63 lines · 62 tokens per session scan A c6a3001c5c2d
site-preship-review is a skill published in the GitHub repository mtarcure/claude-vibe-squad (211 stars, last pushed yesterday), licensed MIT. It adds 62 tokens to every session and 1,144 once invoked, about $0.0002 per session on Opus 5.5. A static security scan graded it A with 1 finding (makes network calls). No closer match exists in the catalogue, so it is treated as the original; first seen 2026-10-29.
Other skills, from other repositories
verify
Exercise the real app/API/CLI and collect observable evidence; tests alone do not count as end-to-end verification.
webapp-testing
Start/reuse a local app, wait for readiness, inspect rendered state/console/network, act from observed selectors, and verify with evidence.
ux-grill
Use when a running web app's UX has to be audited reproducibly rather than by feel — a deterministic mechanical pass over routes and viewports (axe, target size, horizontal overflow, page title, scripted journeys) followed by a screenshot-grounded interrogation of the operator, journey by journey, with two persona…
dogfood
Systematically explore and test a web application to find bugs, UX issues, and other problems. Use when asked to "dogfood", "QA", "exploratory test", "find issues", "bug hunt", "test this app/site/platform", or review the quality of a web application. Produces a structured report with full reproduction evidence -…
ux-review
Verify a shipped UI against objective design gates — WCAG accessibility, design-token fidelity, responsive/render integrity, interaction & state coverage, and visual regression — by rendering it in a real browser. Use for 'ux review', 'a11y check', 'does this UI hold up', or /ux-review. Enforces design quality; does…
pursr
Use Pursr for browser screenshots, scripted visual operation, visual regression, accessibility audits, DOM inspection, and MCP-driven browser sessions. Use when a user asks an agent to inspect a site, operate an existing browser session, fill or draft UI content, record a tutorial, compare visuals, debug layout, or…