What is kernelcad-design-loop?

A workflow for creating or revising 3D CAD models as editable TypeScript source and testing them through a hosted geometry engine. It repeats evaluation and design review until the model is accepted or work must stop.

Build mechanical parts, housings, mechanisms, gearboxes, and robot arms; diagnose failures; review the result in Studio; and prepare it for export.

It gives the agent a structured way to find and repair failed geometry instead of guessing at code or relying on a dry run alone.

kernelcad-design-loop — Skill for Claude Code by w1ne/kernelCAD-web

Install

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.

Any agent

npx skills add w1ne/kernelCAD-web --skill kernelcad-design-loop

Clone the repo

git clone --depth 1 https://github.com/w1ne/kernelCAD-web

Made for: Claude Code.

Or install kernelcad, the plugin that ships this one along with the rest of its 4 skills.

Wrote 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.

README.md

[![agentmods](https://agentmods.dev/badge/skills/w1ne/kernelcad-web/kernelcad-design-loop/github.svg)](https://agentmods.dev/skills/w1ne/kernelcad-web/kernelcad-design-loop)

Your own site

<a href="https://agentmods.dev/skills/w1ne/kernelcad-web/kernelcad-design-loop"><img src="https://agentmods.dev/badge/skills/w1ne/kernelcad-web/kernelcad-design-loop/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.

Your own site · 80×15

<a href="https://agentmods.dev/skills/w1ne/kernelcad-web/kernelcad-design-loop"><img src="https://agentmods.dev/badge/skills/w1ne/kernelcad-web/kernelcad-design-loop.svg" alt="Reviewed on agentmods" width="80" height="20"></a>

Per session 56 Skills are progressive disclosure: only the name and description are preloaded; the body loads when the skill is used.

When invoked 460 The whole file, excluding the scripts and references it only reads on demand.

Security scan A 0 findings. A grade says what 26 rules found in the file — not that it is safe.

Origin original No closer match found in the catalogue.

Token cost

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	$0.00056	$0.00460
Opus 5.5	$0.00022	$0.00184
Sonnet 5.5	$0.00011	$0.00092
Haiku 4.5	$0.00006	$0.00046

Measured 2d ago against content hash 1c991dc780c4, method: parsed. Prices are Anthropic first-party input rates as of 2026-10-07, from the pricing page.

Security

Grade A, and why

kernelcad-design-loop 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.

plugins/kernelcad/skills/kernelcad-design-loop/SKILL.md · 35 lines

What it actually says

kernelCAD design loop (hosted MCP)

The kernelcad MCP server (https://mcp.kernelcad.com/mcp) runs the OpenCASCADE kernel for you. Nothing needs to be installed locally. Discover the live tool list from the server; tool availability depends on whether the user is signed in.

Loop

Check for a canonical pattern first: lookup_cookbook({ query }) (see the kernelcad-cookbooks skill). Start from a hit when one exists.
Write .kcad.ts source. Use lookup_api({}) for the API surface instead of guessing.
Iterate fast with evaluate_script({ code, dryRun: true }), then finish with a full evaluate_script({ code }). A dry run is not evidence that the solid builds.
For housings, mechanisms, gearboxes, robot arms and other production builds, call design_loop({ goal, attempts: [{ code }] }). Apply its nextActionPrompt / revisionAssist patches and repeat until ok: true or convergence.escalate says to stop and ask the user.
On failure, why_did_this_fail({ code }) then repair_script({ code }). lookup_diagnostics({}) explains any diagnostic code.
Show the result once with open_in_studio (see kernelcad-studio-paint), then export (see kernelcad-export).

Rules

Report the numbers the tools return (volumes, bbox, diagnostics); do not describe geometry you have not evaluated.
Keep the source as the deliverable: give the user the final .kcad.ts.
If a tool says sign-in is required, tell the user to complete the OAuth prompt their client shows for the kernelcad server.

Changes

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 · 35 lines · 56 tokens per session scan A 1c991dc780c4

Subscribe to this mod's changes

kernelcad-design-loop is a skill published in the GitHub repository w1ne/kernelCAD-web (30 stars, last pushed today), licensed MIT. It adds 56 tokens to every session and 460 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 2026-10-14.

3d-modeling Design and UX NURBS ai-agents cad mcp model-context-protocol

simplecadapi

Build, assemble, inspect, reconstruct, and export parametric CAD models with the SimpleCADAPI Python SDK. Use for SimpleCAD geometry modeling, constrained sketches, parts and assemblies, standard gears and bearings, STEP/BREP inspection and reconstruction, vector-PDF drawing reconstruction, durable product packages…

NiJingzhe/SimpleCADAPI · 71 tokens

aieng-cad-authoring

Active MCP-first CAD authoring skill for AIENG Workbench. Use when the user asks to create, design, draft, model, or iteratively build CAD geometry through the active cad. MCP tools. Do not use for solver/result review, schema/library implementation, or editing legacy IR plans.

armpro24-blip/cad-cae-copilot · 67 tokens

aieng-closed-loop-copilot

MCP-first closed-loop CAD/CAE copilot skill for improving an existing solved design against explicit targets. Use when the user asks for bounded design iteration using editable CAD parameters, solver reruns, and evidence-backed comparison.

armpro24-blip/cad-cae-copilot · 54 tokens

rhino

Automate Rhino modeling through RhinoCommon, work with Grasshopper data trees, and configure Rhino.Compute workflows.

alivirgo/Major-AI-Skills · 24 tokens

mechanical-drafting-gbt

Create, revise, audit, validate, and present mechanical CAD using Chinese GB/T drafting conventions. Use for AutoCAD/AutoCAD MCP; parametric 2D/3D parts, complex assemblies, mechanisms, engines, gears, machined parts, consumer products, enclosures, and powered appliances; DXF/DWG/PDF/STEP and model-based definition…

beiming183-cloud/AutoCAD-skills · 145 tokens

modulor

2D drafting + 3D modeling via Modulor (agent-native CAD). Use when the user asks to draw floor plans, model buildings or parts, produce DXF/SVG drawings, STL/GLB meshes, IFC BIM models, or to import and modify existing DXF drawings. Requires pip install modulor.

bcllcc/modulor · 69 tokens