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 skills/autonomous-ai/autonomous-workshop/design-referencenpx skills add autonomous-ai/autonomous-workshop --skill design-referencegit clone --depth 1 https://github.com/autonomous-ai/autonomous-workshopWhat 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 | $0.00066 | $0.00903 |
| Opus 5 | $0.00033 | $0.00451 |
| Sonnet 5 | $0.00013 | $0.00181 |
| Haiku 4.5 | $0.00007 | $0.00090 |
Grade A, and why
design-reference scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 2d ago.
A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.
Nothing flagged
None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.
How it starts
The opening of the file, as written. The whole thing — 84 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Design references
Research the Internet before selecting a construction design. Use authoritative sources for specifications and licensed analogous designs for construction patterns; never replace the user's reference image, dimensions or requested exterior with the nearest-looking result.
Boundary
- Use this skill for shape archetypes, feature ordering, parametric CAD idioms, mechanism construction examples, and sourced design constraints.
- Use manufacturer documentation, standards, or an official technical source for numerical specifications. A third-party analogous model is not authority for the user's dimensions, ratings or scale.
- Use
step-partsfor a bought motor, bearing, fastener, board, connector or other standard component the model must physically fit. Download its canonical STEP into<project-dir>/ref/, derive the seat with$cad'scadmount, declare it inmeasure/mounts.json, and runcheck_mount. - Use
$electromechanical-integrationwhen the question includes power, control, wiring, electrical ratings, exact lamp/socket contacts or removable powered interfaces. A design repository is not electrical authority. - A design reference is not proof of likeness, manufacturability, assembly or physical fit. Those claims remain with the project gates.
Internet research workflow
- Read and measure the user's evidence first. Name the one or two exact design questions research must answer; do not browse with only a product-category query.
- Write a research contract for each question: required facts, acceptable source authority, geometry or operation uncertainty, packaging constraints, license needs, and the condition that would reject a candidate.
- Search the Internet with form, feature and operation terms. For example,
search
tapered shell vent slots parametric CAD, not merelyenclosure. - Review no more than five strong candidates. Prefer, in order:
- manufacturer documents, standards and official technical pages for dimensions, ratings and compatibility facts;
- repositories with source CAD or build123d/CadQuery/OpenSCAD code, an explicit license, and an immutable commit or release for construction patterns;
- authoritative CAD/kernel documentation for operation-specific questions.
- Record every used result and meaningful rejection: query, stable URL,
repository plus commit/release when available, license/use, exact claim or
specification taken, relevant feature, construction lesson, and status
(
used,rejected,miss, orunavailable). - Compare the evidence against the research contract, select the design, and re-measure every user-specific dimension from the user's image/spec. Do not copy an analogous model's placement, scale or silhouette.
What ships with it
1 file 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.
- 2d ago First seen · 84 lines · 66 tokens per session scan A cb00f5f7064a
design-reference is a skill published in the GitHub repository autonomous-ai/autonomous-workshop (11 stars, last pushed 2d ago), licensed Apache-2.0. It adds 66 tokens to every session and 903 once invoked, about $0.0003 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
variant-design
AI-driven interactive design generation, style analysis, and UX evaluation with Impeccable design system. Context-aware output — auto-detects React projects (package.json with react dependency) and generates .tsx components; otherwise generates zero-dependency interactive HTML. Seven modes — (0) Design System: tokens…
variant-generate
Generate 3 distinct design variations from a prompt. Each variation feels like a different studio — layout, palette, typography, and motion diverge. Supports variation actions: Vary strong/subtle, Distill, Shuffle layout, Change style, Remix colors, Mix, Dramatize, Add motion, Make interactive, Polish, Critique, See…
variant-design-system
Design System mode — tokens → components → pages three-layer workflow. Generate a complete token set (palette/type/spacing/motion/elevation/radius), confirm it, then all downstream outputs are visually consistent. Triggers on: ds, design system, tokens, create token set, define palette, define tokens, set up tokens…
variant-ux
UX Review mode — heuristic evaluation (Nielsen's 10), cognitive load analysis, mental model diagnosis, affordance audit, dark pattern detection — grounded in NNG research. Load references/ux-heuristics.md and references/ux-psychology.md. Triggers on: ux review, heuristic evaluation, usability audit, cognitive load…
create
Design and build a finished motion piece (default vertical 1080×1920; landscape/square also supported) from a brief through the 甲乙环 (critic-loop) pipeline — a unified design+build agent (乙) drafts a design and writes the Remotion code in one continuous context; N independent draws are blind-selected for the most…
variant-analyze
Audit existing sites, extract design tokens, generate style-matched pages. Runs consistency checks and Visual Quality Audit (40-item anti-slop checklist, Quality Score XX/100). Triggers on: analyze my site, audit this, match this style, extract tokens, what's wrong with this design, migrate, add a page to my site…