- Home
- Skills
- Agents & Orchestration
- solidcraft modeling
Works with the AI Agents you already use
solidcraft modeling
A systematic MCP workflow for building, batching, and verifying parametric 3D CAD models in SolidCraft.
Free
solidcraft modeling
Example session with this skill installed
I need a parametric mounting plate, 100x80x5mm, with four M4 holes at the corners (10mm offset) and 5mm fillets on the outer corners. Build this in SolidCraft.
- Read your context and instructions
- Compiled the solidcraft modeling
I have planned and executed the parametric mounting plate build in two batches to ensure structural integrity and correct feature ordering.
SolidCraft Build Summary
| Stage | Operations | Status | Metrics |
|---|---|---|---|
| 1. Base Geometry | Parameters, Sketch, Extrude | Success | Vol: 40000mm³, BBox: 100x80x5 |
| 2. Finishing | Fillets, Hole Pattern | Success | Vol: 39547mm³, Faces: 18 |
// Verification of Center of Mass (COM) and Volume
// Expected Volume: 39547.43 mm^3 (after fillets and holes)
// Result: 39547.43 mm^3 (Match)
// Batch 1: Parameters and Base Plate
// - Set w=100, d=80, h=5
// - Rectangular sketch centered on origin
// - Extrude base feature
// Batch 2: Refinements
// - Apply 5mm fillets to vertical corner edges
// - Create M4 hole sketch on top face
// - Rectangular pattern for 4 corner holes with expression spacing [w-20, d-20]
Next steps
- Run
set_parameter {"name": "w", "value": 120}to verify the holes follow the plate edges correctly. - Execute
export {"path": "mounting_plate.step"}to generate the CAD file. - Call
screenshot {"source": "model", "view": "iso"}for final visual QA.
Connects securely to your tools. The creator never sees your data.
About this skill
Building CAD models through an MCP server often results in excessive token usage and fragile, incremental builds. This skill provides a systematic workflow for SolidCraft modeling, reducing call volume by grouping operations into strategic batches.
What it does
- Batch execution groups parameters, sketches, and features into single-stage calls to minimize API overhead.
- Predictable geometry uses origin-centered placement and expression-based dimensions for reliable patterns and mirrors.
- Validation loops verify every stage using
measurecalls to check volume, bounding boxes, and center of mass against computed values. - Failure recovery handles all-or-nothing batch failures by identifying specific indices and targeting fixes to the failed step.
- Visual verification generates isometric and top-view screenshots to confirm feature placement that numerical measures might miss.
How it works
- Initial inspection runs
inspect_designto check connection status and existing features before planning. - Phase modeling structures the build into three distinct batches: base features, finishing (fillets/shells), and secondary operations (holes/patterns).
- Rigorous verification compares post-call metrics like COM shift and face counts against pre-computed geometry targets.
Frameworks & tools
SolidCraft CAD software, MCP (Model Context Protocol) servers, and standard export formats including .step, .3mf, .stl, and .urdf.
Why this beats prompting it yourself
Standard prompting often leads to one-tool-call-at-a-time workflows that waste tokens and frequently lose track of sketch IDs. This skill uses a hardened reference for ID prediction and batching strategies that prevent timeline errors and ensure parameter-driven stability.
Use cases
- Parametric part design where dimensions must update via variables without breaking the feature tree.
- 3D printing preparation requiring shelled enclosures with specific wall thicknesses and fillet orders.
- Robotics simulation for generating assemblies and exporting valid .urdf files for ROS or Mujoco.
- Headless CAD automation where designs are built and exported in background environments without a GUI.
Known limitations
Mirrored or patterned features cannot be mirrored or patterned again. Boolean operations between touching curved faces are high-risk and require isolated batching.
Mirrors and patterns cannot be nested. Booleans on touching curved faces require isolated batches to avoid rollback. Body topology data is heavy and splits cylinders into half-faces.
How to use
- 1
Connect your agent to the Agensi MCP
Copy the prompt and paste it into Claude, ChatGPT, Cursor or any agent that supports MCP. It sets up the connection for you. Manual setup
- 2
Sign in to Agensi
Your agent shows a sign-in link. Sign in and click Allow.
- 3
Ask for what you need
Your agent loads this skill and uses it on your task. It always asks before any payment.
Reviews
No reviews yet
Be one of the first to try it. Every listed skill goes through the trust checks below.
Security scanned
Checked by our 8-point scan before listing
1 use
Times used to date
Free forever
No account required to browse
Trust & safety
Security scan
Security scan passed 19 days ago
- Free with an Agensi account