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 Nutlope/variate --skill variategit clone --depth 1 https://github.com/Nutlope/variateWrote 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/nutlope/variate/variate)<a href="https://agentmods.dev/skills/nutlope/variate/variate"><img src="https://agentmods.dev/badge/skills/nutlope/variate/variate/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/nutlope/variate/variate"><img src="https://agentmods.dev/badge/skills/nutlope/variate/variate.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 | $0.00168 | $0.03646 |
| Opus 5.5 | $0.00067 | $0.01458 |
| Sonnet 5.5 | $0.00034 | $0.00729 |
| Haiku 4.5 | $0.00017 | $0.00365 |
Grade A, and why
variate 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 — 267 lines — stays where its author put it; the contents beside it link to each section on GitHub.
variate
The user's own localhost is the canvas. You write real alternatives of one of their files; a small card at the bottom of their page lets them flip between them; the one they keep is the code. There is no studio, no preview server of ours, and no separate thing to keep in sync.
Critical floor
Obey these even if you read nothing else.
- Variant 1 is the user's file as it was. Never edit it, never delete it, never overwrite it. Everything you write is a new numbered sibling.
- Every variant is a complete, drop-in replacement for the target file. Same exports, same props, same imports the app relies on. It must run.
- Speak the project's own language. Its design tokens, its utility classes, its components, its fonts, its real copy. A variant that adds a dependency or invents a colour has already failed.
- Four positions that disagree. Each changes something different, and
plan.jsonsays what and what it costs. Four tweaked card grids is wallpaper, not a round. Writeplan.jsonbefore the variants;checklints it. - A round narrows, it never accumulates. Once the user favours one, the
others have done their job:
variate narrowbefore drawing again, so they choose between live options rather than re-reading rejected ones. It hides rather than destroys, in.variate/<set>/.dropped/, andstatuslists them, so "go back to the split one" is a copy, not a redraw. - Talk in positions, never filenames. "2 of 4", not "3.tsx".
- Ask one question per round, with your recommendation.
- Presenting a round never ends your turn. Hand it over, then run the listening loop below. The card's keep, refine and pick land there.
- Drain before you answer, every turn a round is open. The user decides
on their own clock and may have clicked an hour ago, so
variate drainis the first thing you run in any project with a.variate/directory, even when their message has nothing to do with design. An ask left sitting is the user waiting for an answer you already have. - Never write the target file directly while a set is open. Write a variant and switch to it. The user's own hand edits are sacred: the next switch adopts them as a new variant, never destroys them.
variate endwhen the session is done.end <set>closes one round and keeps its live file; the barevariate endcloses the session: it keeps what is live, deletes the rest, removes the tag, and stops variate's own little server.
What ships with it
6 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 First seen · 267 lines · 168 tokens per session scan A 19e3fba5e7c3
variate is a skill published in the GitHub repository Nutlope/variate (212 stars, last pushed 1mo ago), licensed MIT. It adds 168 tokens to every session and 3,646 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
webgl-holographic-foil
A self-contained WebGL2 hero: thin-film interference over a crushed-foil surface whose palette shifts with the viewing angle; move the cursor to tilt the film.
html-ppt-hermes-cyber-terminal
OpenDesign + BYOK: choosing and wiring your own model, hands-on — cost, quality, and the routing decision. Built as a decision-grade AI literacy deck for engineers, IT, applied-AI teams.
html-ppt-taste-brutalist
16:9 HTML deck in tactical-telemetry / CRT-terminal taste. Deactivated-CRT charcoal slides, white-phosphor monospace, hazard-red accent, scanline overlay, ASCII syntax, density over decoration. Distilled from Leonxlnx/taste-skill brutalist-skill (Tactical Telemetry mode).
visual-ralph
Visual Ralph orchestration for frontend UI from generated references, static references, or live URL targets, using $ultragoal with built-in visual verdict and pixel-diff evidence until the implementation matches and leaves a reproducible design system.
glance-canonical-layouts
Use this skill to create home screen widgets using Jetpack Glance Canonical Layouts from the Android Platform Samples project, or when asked to implement standard widget layout patterns (including Check List, Image Grid, Snap Scrolling, and Standard Toolbar).
accessibility
Consolidated accessibility skill entrypoint for WCAG 2.2, ARIA Authoring Practices, cognitive accessibility, Section 508, EN 301 549, design intent verification, and the Accessibility Planner workflow.