Orkas is a desktop application for commanding a team of AI agents through one chat, with a commander model assigning work to specialist agents in parallel or in sequence. People use it to coordinate research, writing, presentations, and software tasks while keeping files on their computer. The catalogue includes skills for extending the agents available to Orkas.
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 Orkas-AI/Orkas --skill ui-craft-checksgit clone --depth 1 https://github.com/Orkas-AI/OrkasWrote 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/orkas-ai/orkas/ui-craft-checks)<a href="https://agentmods.dev/skills/orkas-ai/orkas/ui-craft-checks"><img src="https://agentmods.dev/badge/skills/orkas-ai/orkas/ui-craft-checks/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/orkas-ai/orkas/ui-craft-checks"><img src="https://agentmods.dev/badge/skills/orkas-ai/orkas/ui-craft-checks.svg" alt="Reviewed on agentmods" width="80" height="20"></a>- NVIDIA SkillSpector pass
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 | $0.00005 | $0.02213 |
| Opus 5 | $0.00003 | $0.01107 |
| Sonnet 5 | $0.00001 | $0.00443 |
| Haiku 4.5 | $0.00001 | $0.00221 |
Grade A, and why
ui-craft-checks 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 13d 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 — 128 lines — stays where its author put it; the contents beside it link to each section on GitHub.
ui-craft-checks
Use this comprehensive gate for formal review, QA, launch handoff, exact-fidelity inspection, high-risk complex UI, or when the fast gate exposes a deeper craft problem. Ordinary single-screen work uses ui-design-executor and its bundled validator without loading this full matrix. Pair standalone artifacts with ui-artifact-workspace when this specialist is triggered.
This skill does not replace ui-html-renderer or ui-design-review. It tells them what must be checked before delivery.
Run the gate internally by default. Surface the full checklist only for review, handoff, QA, or when the user asks for the detailed pass/fail matrix.
When To Use
Use this skill when:
- The user asks for a formal review, QA matrix, hardening, launch-readiness, accessibility audit, or detailed handoff.
- Exact screenshot/design-source fidelity needs comprehensive validation.
- A complex form, live/data surface, or composite interaction is high risk or has failed the fast gate.
- The user explicitly requests a full craft or anti-template critique.
For ordinary single-page delivery and tiny text-only answers, skip this skill.
Severity
P0: Must fix before final delivery.P1: Should fix before production or stakeholder review.P2: Polish; fix when scope allows.
P0 Gates
The artifact fails the craft pass if any P0 is present:
- Not HTML by default: a UI/design deliverable is only prose, PNG, SVG, Markdown, or a raw asset gallery when the user did not request that final format.
- Explicit format ignored: the user requested a non-HTML final format but the artifact silently substitutes HTML or another format for it.
- Detached revision: a normal follow-up creates a new sibling or regenerates an unrelated artifact instead of inspecting and patching the canonical artifact in place, losing accepted content or user edits.
- Broken artifact boundary: a standalone deliverable has no stable artifact directory or valid manifest/entry, or it depends on missing local files, machine-specific absolute paths, undeclared secrets, or an unmentioned build step that prevents sharing or packaging.
- Source drift: screenshot, Figma/design-source, or explicit contract structure is replaced by an unrelated dashboard, table, chart, sidebar, pricing page, or kanban board.
- Missing critical states: a data-fetching or data-transforming surface does not implement reachable loading, populated, empty, error, and edge/partial states. Mentioning them in rationale while shipping only populated DOM still fails.
- Broken form validation: required fields lack labels/helper/error wiring, or the prototype omits pristine, dirty/touched invalid, submitted-pending, recoverable server-error, or success behavior that the workflow can reach. A user-requested success/failure outcome that exists only in an untriggered branch still fails.
- Keyboard/accessibility failure: interactive controls are not keyboard reachable, focus-visible is absent, icon-only controls lack names, or native elements are replaced with inert divs.
- Unsafe live data: live-ready artifacts persist tokens, credentials, cookies, auth headers, raw provider responses, or secret-like fields.
- Text/layout collision: text overlaps controls, is clipped without intent, or mobile layout hides the primary action.
- Broken HTML/runtime: invalid or malformed HTML, unclosed critical tags, inline JavaScript syntax/runtime errors, or missing root content causes a blank or unusable first render. Inline event handlers nested inside
innerHTMLstrings are a P0 because their quote/escape layers are fragile. Mutating cached element-reference maps to redirect a renderer, clearing a container before querying/moving its former descendants, or wrapping onlyaddEventListener("DOMContentLoaded", init)intry/catchwhile leaving the later callback unguarded is also P0; use immutable references, detached fragments, one safe commit, a genuinely guarded initializer, and data hooks plus listeners/delegation. - False capability claim: the response claims Figma import, connector refresh, external-browser validation, or accessibility checks happened when they did not.
- Invented evidence: sample metrics, placeholder records, inferred copy, or guessed assets are presented as observed source content or production data.
- Raster route bypassed: a requested original raster asset is replaced by an inline SVG/gradient or the whole UI is delegated, even though the dedicated raster capability is available. The asset brief also fails if its stated aspect ratio conflicts with its pixel dimensions, or if it names a save path without wiring that relative raster path and an honest fallback into the HTML.
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.
- 13d ago First seen · 128 lines · 5 tokens per session scan A e0bba6a3597e
ui-craft-checks is a skill published in the GitHub repository Orkas-AI/Orkas (1,911 stars, last pushed yesterday), licensed MIT. It adds 5 tokens to every session and 2,213 once invoked, about $0.0000 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 2026-08-30.
Other skills, from other repositories
tutti-ui-system
Use when working with @tutti-os/ui-system components, replacing local UI with shared components, querying component ids or metadata, promoting UI into shared base or business components, or maintaining UI-system storyboard inventory.
cross-border-listing-image-layout-audit
Audit listing image and layout notes — first-impression hero, info hierarchy, lifestyle vs studio, copy density, trust signals, compliance — into a follow-up checklist.
design-plan-review
Reviews a UI or UX plan before implementation.
excalidraw-diagram-generator
Converts textual concepts or workflows into clear diagram instructions suitable for Excalidraw or other visual tools.
infographic-builder
Turns textual content into structured infographic formats suitable for reports, presentations, and educational materials.
ui-ux-layout-advisor
Advises on interface layouts to optimize clarity, spacing, hierarchy, and usability.