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 open-octo/octo-agent --skill artifact-designgit clone --depth 1 https://github.com/open-octo/octo-agentWrote 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/open-octo/octo-agent/artifact-design)<a href="https://agentmods.dev/skills/open-octo/octo-agent/artifact-design"><img src="https://agentmods.dev/badge/skills/open-octo/octo-agent/artifact-design/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/open-octo/octo-agent/artifact-design"><img src="https://agentmods.dev/badge/skills/open-octo/octo-agent/artifact-design.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 | $0.00187 | $0.02808 |
| Opus 5.5 | $0.00075 | $0.01123 |
| Sonnet 5.5 | $0.00037 | $0.00562 |
| Haiku 4.5 | $0.00019 | $0.00281 |
Grade A, and why
artifact-design 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 1mo 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 — 184 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Artifact design
An artifact is any .html/.htm/.md/.markdown/.png/.jpg/.jpeg/.gif/.svg/.webp
file the agent produces. Writing one through write_file/edit_file surfaces it
automatically in the web UI's Artifacts panel; a file built some other way (a
script, a build step, a download) needs one show_artifact call with its
absolute path. This skill is about what to put inside the HTML — read it
before writing the first line.
If the page contains any chart, graph, plot, heatmap, sparkline, or stat
tile, read references/charts.md before writing the chart — chart-type
selection, the color system (with a validated default palette in
references/palette.md), legend/axis/tooltip conventions, and chart
legibility at the panel's docked width.
How the panel actually works — design within these constraints
- The page is a real web page. HTML renders inside a frame from a path on
octo's own address, so everything a normal web page can do works:
localStoragepersists (the page's keys are kept apart from octo's own and from other pages'),<a download>saves a file,requestFullscreen()and pointer lock work, WebGL and Web Audio work, andfetchcan reach public APIs. Do not write code that calls octo's own API from inside the page. - Files beside the page load by relative path.
<script src="./app.js">,<link href="./style.css">,<img src="./chart.png">,loader.load('./model.glb'),fetch('./data.json'), fonts, audio, video and other.htmlpages in the same directory (or a subdirectory) are served with the page. Never start such a path with/: the page lives under a path prefix, and an absolute path lands outside it. A single file is still the simplest artifact; split into sibling files when the page has a real script or a binary asset (a model, a font, a recording) that would be absurd to inline. - External resources load from anywhere — which is why to be sparing. Pin
exact versions; if the user is in mainland China, prefer a mirror reachable
there (
cdn.bootcdn.net,cdn.staticfile.net,registry.npmmirror.com). Reach for a CDN only when the page needs a real library (React, ECharts, Chart.js, three.js, …) — a page that depends on one shows nothing when that host is unreachable, and a saved Light App has to keep working for years. - The default viewport is narrow. The panel is a 420px-wide docked
sidebar by default; the user can maximize it to
min(900px, 75vw), but don't design for that as the common case. Build the layout to read cleanly at ~380–420px first, then let it use extra space gracefully above that — not the other way around. This is the opposite of most artifact platforms, where the canvas starts wide. Multi-column layouts, wide tables, and side-by-side diagram lanes need an explicit@media (max-width: 720px)(or tighter) fallback to a single column, or they'll clip or force horizontal scroll in the default view. - The theme arrives on the URL — read it. octo loads the page with
?theme=darkor?theme=light: the palette the user picked in the Web UI, which is not necessarily the OS's. Nothing applies it for you, so a page that only watchesprefers-color-schemefollows the OS and becomes the one dark rectangle on a light screen the moment the two disagree. Apply it at startup and key the palette offdata-theme, keeping the media query as the fallback for a page opened bare in a browser tab:
What ships with it
2 files beside SKILL.md in the same directory: the scripts, references and assets a skill reads on demand. Not counted in the per-session cost; read them before you install if any of them is executable.
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.
- 1mo ago Changed · -11 lines · -8 tokens per session 5d6d650b0c14
- 1mo ago Changed · +22 lines 06afdd3618d0
- 1mo ago Changed · +18 lines · +13 tokens per session cd15c8455e66
- 1mo ago Changed · +6 lines · +7 tokens per session 91ec82d5d913
- 1mo ago First seen · 149 lines · 175 tokens per session scan A 9c3347dbe871
artifact-design is a skill published in the GitHub repository open-octo/octo-agent (138 stars, last pushed today), licensed MIT. It adds 187 tokens to every session and 2,808 once invoked, about $0.0007 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 2026-08-30.
Other skills, from other repositories
mermaid-diagram
Use when the user asks for a flowchart, architecture diagram, sequence diagram, ER diagram, state machine, gantt chart, or any visual that fits Mermaid syntax. Hope Agent's chat renders Mermaid blocks natively, so the diagram appears inline.
claude-design
Design one-off HTML artifacts (landing, deck, prototype).
sketch
Throwaway HTML mockups: 2-3 design variants to compare.
design-md
Author/validate/export Google's DESIGN.md token spec files.
baoyu-infographic
A generator for infographics, which are visual summaries that organize information with text, shapes, and relationships. It combines one of 21 information layouts with one of 21 visual styles and supports custom sizes and languages.
architecture-diagram
Dark-themed SVG architecture/cloud/infra diagrams as HTML.