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 agentmods add commands/geo-mena/distill/articlegit clone --depth 1 https://github.com/geo-mena/distillWhat 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 | $0.00025 | $0.00979 |
| Opus 5 | $0.00013 | $0.00490 |
| Sonnet 5 | $0.00005 | $0.00196 |
| Haiku 4.5 | $0.00003 | $0.00098 |
Grade A, and why
article 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 2d 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.
What it actually says
Load the distill-design skill, then produce a long-form Distill-style article on the topic given in $@.
Workflow:
- Read
plugins/distill-design/SKILL.mdfirst for the file table and the decision tree, thenplugins/distill-design/DESIGN-SYSTEM.md§ "Content fundamentals" and § "Typography" for voice and exact display sizes. - If the topic is ambiguous (paper not named, audience unstated, length unstated), ask one clarifying question. Otherwise commit.
- Outline the article first as a numbered list of h2 sections in sentence case, with one-line descriptions. Pause for confirmation only if the user asked you to. Otherwise proceed.
- Copy
plugins/distill-design/ui_kits/article/index.htmlas the scaffold. Replace prose, citations, and diagrams; keep the<head>, the Babel-standalone setup, the<link>to../../tokens/colors_and_type.css, and the<script src>lines forPrimitives.tsx,Chrome.tsx,Diagrams.tsx, and any ofRL.tsx/Graph.tsx/Heatmap.tsx/Specialized.tsxyou need. - Use
ArticleNav,ArticleHeader,TOCDrawer, andArticleFooterfromChrome.tsx. The footer must include all 7 canonical sections (Acknowledgments, Author Contributions, Discussion and Review, References, Updates and Corrections, Reuse, Citation) — each is a prop onArticleFooter. Pass meaningful values; do not leave the defaults. - Inline citations use the
Citationprimitive fromPrimitives.tsxwith abibmap. Equations useMathBlock. Figures useFigurewithn(figure number) andcaption. Diagrams inside figures compose primitives — pick the right file via the SKILL decision tree (RL →RL.tsx, beam search / HMM →Graph.tsx, attention / convolution / recurrent →Heatmap.tsx, molecule / Feynman / annotated raster →Specialized.tsx, generic FiLM-style →Diagrams.tsx+Primitives.tsx). - Default output location: a sibling HTML file under
plugins/distill-design/preview/article-<slug>.html. If the user asks for a TSX-only scaffold to drop into their own app, return a.tsxfile with the article body component (no<html>shell) and list the imports they need.
Voice and prose constraints (non-negotiable):
- English only. Never translate the artifact, even if this conversation is in another language.
- First-person plural ("we") for narration; second person ("you") only when addressing reader actions.
- Active voice, present tense. Math verbs get verbed ("dot producted", "softmaxed").
- Sentence case for headings. Title case reserved for proper nouns and named methods.
- No emoji anywhere. No exclamation points in body. No contractions in formal sections.
- No marketing superlatives ("revolutionary", "powerful", "robust", "comprehensive", "elegant", "game-changing"). State what the thing does; let the reader judge.
- No year literals in the article body unless the topic genuinely requires dating a specific paper. Prefer "now-archived", "modern", "earlier".
- Captions describe what to look at, not what to conclude.
Visual constraints (non-negotiable):
- Body color
rgba(0,0,0,0.8)(≈#333). Pure#000only on the h1. - Reading column 704px; figures may break out to 984px via the
breakoutprop onFigure. - Strokes 1.5–2px in
#666withstroke-linecap="round". No multi-color icons. No gradients, no shadows on body content, no fixed elements. - Background
#fdfdfd. Never pure white.
Verification:
- After writing, serve via
cd plugins/distill-design && python3 -m http.server 8765(Babel standalone fetches the.tsxfiles via XHR —file://is blocked by CORS). - Open
http://localhost:8765/preview/article-<slug>.htmland verify visually withchrome-devtools(screenshot, list-network-requests). The user values verification; do not claim the article is correct without inspecting the rendered output.
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.
- 2d ago First seen · 38 lines · 25 tokens per session scan A ec4cb308f7b4
article is a command published in the GitHub repository geo-mena/distill (2 stars, last pushed 3mo ago), licensed MIT. It adds 25 tokens to every session and 979 once invoked, about $0.0001 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 2026-08-31.
Other commands, from other repositories
design-chatbot
Design a complete chatbot or conversational assistant UI — dialog flows, message bubbles, quick replies, typing indicators, error states, and accessibility.
ux-recommend
The v2 flagship. Runs the 5-parallel-search Python engine across 11 data manifests (75+ styles, 170+ palettes, 65+ type pairs, 120+ components, 170+ industries, 110+ UX laws, 50+ motion presets, 120+ anti-patterns, 92 brand specs) and returns a complete recommended design system for your project — with always-on…
init
Initialize a project for on-brand Marmo UI generation — detect or interview for brand basics (layout, primary color, fonts, radius, spacing) and write a DESIGN.md if one doesn't exist.
database-setup
Use when a project needs to store data and has no database yet. Setting up Supabase, creating tables, writing queries, and connecting them to the frontend. Written for designers.
review
Cold re-quiz on code that already shipped — your own session commits, not the change in front of you.
design-critic
Use when you want an honest read on a design instead of encouragement. What is wrong, ranked by severity, with the reason attached. Works on a screenshot, a mockup, or a described screen.