Ouroboros is a general-purpose AI agent that maintains its identity, memory, and task history across restarts while working on external projects and coordinating specialist agents. It runs as a desktop application or headless command-line tool, using configured remote APIs or a local model for inference. The catalogue skills support working with or modifying Ouroboros.
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 razzant/ouroboros --skill design_demogit clone --depth 1 https://github.com/razzant/ouroborosWrote 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/razzant/ouroboros/design_demo)<a href="https://agentmods.dev/skills/razzant/ouroboros/design_demo"><img src="https://agentmods.dev/badge/skills/razzant/ouroboros/design_demo/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/razzant/ouroboros/design_demo"><img src="https://agentmods.dev/badge/skills/razzant/ouroboros/design_demo.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 | $0.00040 | $0.00870 |
| Opus 5.5 | $0.00016 | $0.00348 |
| Sonnet 5.5 | $0.00008 | $0.00174 |
| Haiku 4.5 | $0.00004 | $0.00087 |
Grade A, and why
design_demo 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 today.
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 — 47 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Showing a visual choice
Write this sentence first: "On this page you compare X and decide Y." It becomes the page title and the first line of any brief. The person will spend a minute: a gallery of every state is an archive and a folder of screenshots is evidence; neither is a demo.
A change to an approved look is decided on renders of both versions, never on a sentence. If a variant moves away from something the owner already approved, show it and say so.
The page
- One self-contained file: the chat's Open copies only that file, so no server, command, sibling asset or URL parameter.
assets/compare.htmlis the template; frames go inline. - One scene at a time: Now and the variants, swapped in place by the strip or keys 1–9 and arrows. No fade, no reload; scroll kept, the compared element anchored. A blink between two pictures hides the difference being judged.
- Real size, 1:1, at the viewer's window. Never shrink windows to fit several; side by side is a second view showing the compared element at 1:1.
- Real product code and the real long data (ui_design §1).
- One plain line per variant: what changes, what it costs. A is the recommendation, chosen for fit over novelty. Always "none of these, here is what bothers me".
- Demo chrome stays outside the product frame; nothing floats over the app. Limitations go in the message, not on the page, unless they change the decision.
The variants
- Two to four genuinely different directions, only when the task is a choice; a refinement shows Now and one proposal. Each is a step from today, named by what the viewer sees change.
- One direction per parallel designer; a set that converged on one composition is rejected before anyone sees it.
- A model critic may drop clearly worse or broken variants; close calls go to the owner.
- A motion variant plays by itself and loops, respecting reduced motion. When the template is the wrong form (motion, a flow), keep its invariants: one file, real size, in-place switch, A recommended, questions on the page.
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.
- today First seen · 47 lines · 40 tokens per session scan A e7b3d8dac3fd
design_demo is a skill published in the GitHub repository razzant/ouroboros (1,572 stars, last pushed today), licensed MIT. It adds 40 tokens to every session and 870 once invoked, about $0.0002 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 2026-10-22.
Other skills, from other repositories
cuga-github-issues
Create GitHub issues for cuga-agent (bugs, features, epics, designs, and related work) against origin using gh, with epic → feature → issue hierarchy and GraphQL sub-issue linking. Use when the user says create new issues, report a bug, file a bug, feature request, new feature, create an epic, GitHub issue, sub-issue…
cuga-contributor-workflows
Run CUGA contributor workflows: Conventional Commits, create pull requests against origin with gh and PR templates, and ruff check/format via uv. Use when the user says commit, create a commit, create PR, pull request, open a PR, ruff check, ruff format, or conventional commits.
data-analysis
Analyze, explore, clean, and visualize datasets with statistical rigor. Use when user asks to analyze data, find patterns, compute statistics, create visualizations, clean messy data, or explore a dataset. Trigger when user says things like "analyze this data", "what trends do you see", "find patterns in", "create a…
web-design
Use when building any web page, HTML artifact, landing page, dashboard, slide deck, report, email, or UI component the user will look at. A taste-first design system that makes generated pages look intentional and crafted instead of generic "AI-made". Provides an aesthetic-direction process, design tokens, a starter…
summarization
Summarize documents, articles, conversations, code, and technical content into concise, accurate summaries. Use when user asks to summarize, condense, create a TL;DR, write an executive summary, extract key points, or distill content. Trigger when user says things like "summarize this", "give me the key points"…
experiment_management
Set up and manage the experiment folder structure. This is Phase 0 — it runs before any analysis begins. All bookkeeping files are JSON (never markdown).