earthtojake/text-to-cad
12 skills50.8k installs132k starsGitHub
Install
npx skills add https://github.com/earthtojake/text-to-cadSkills in this repo
1Cadtext-to-cad is a Python-backed CAD generation and inspection toolkit that converts natural-language requirements, reference images, and 2D technical drawings into validated STEP-first parametric models via build123d source code. Developers use it to author mechanical parts, enclosures, fixtures, and assemblies with named parameters, source-level joints, and explicit positioning; the workflow emphasizes editing Python source rather than generated artifacts. Core operations include STEP/STP generation from build123d Python, direct STEP/STP imports, geometry inspection with selector refs and measurements, validation snapshots, and secondary exports to STL, 3MF, and native GLB. Assembly work leverages AssemblyHelper with named mating datums and build123d joints for functional relationships. --- name: cad description: Create, modify, inspect, and validate STEP-first parametric CAD parts and assemblies. Use for natural-language CAD specs, reference images, 2D technical drawings, STEP/STP generation or direct inspection, Python CAD source, source-level joints, selector references, geometry facts, measurements, mating deltas, snapshots, and secondary STL/3MF/native GLB outputs from CAD ge.6.8kinstalls2Step PartsThe step-parts skill finds and downloads common purchasable CAD parts from the hosted step.parts catalog before creating placeholder geometry. Use https://api.step.parts as canonical API origin. Workflow interprets parts into q search tokens plus optional category, family, standard, or tag facets, searches /v1/parts, presents ambiguous matches, prefers exact actuator STEP downloads when found, and verifies sha256 checksums on download. Network failures are inconclusive: retry with network permission before reporting unavailable. Named servos and motors need alias retries like STS3215 versus ST3215 and family facets. Bundled scripts/download_step_part.py supports search, --id download, facet filters, --all bulk download, and JSON stdout. After creating or updating local STEP files, always hand explicit paths to cad-viewer when installed. API reference in references/step-parts-api.md covers parts index, schema, and OpenAPI. Search guidance notes ANDed query tokens, ORed facet values, and millimeter attributes like thread and lengthMm.5.4kinstalls3SdfSDF is a code generator and validator for SDFormat - the XML schema that describes simulator worlds, robot models, links, joints, poses, sensors, and physics. Developers use it when exporting CAD or robot descriptions to Gazebo or other physics simulators, or when authoring programmatic robot/world definitions via Python `gen_sdf()` functions. The skill manages the full workflow: reading design ledgers, editing generator source (not raw XML), regenerating targets via `python scripts/sdf`, running bundled and optional `gz sdf --check` validation, and handing off completed `.sdf` files to the CAD Viewer. It enforces SI units, frame semantics, design assumptions, and simulator-specific compatibility checks. Generate `.sdf` files from Python `gen_sdf()` sources or edit SDFormat XML with design ledger tracking Validate SDFormat syntax, frame semantics, poses, joints, inertials, and simulator metadata before handoff Support Gazebo5kinstalls4UrdfURDF skill generates and validates robot description files from Python sources, treating URDF work as constrained kinematic modeling rather than simple XML writing. Developers use it when creating or modifying robot links, joints, geometry, inertials, mesh references, and frame conventions - establishing frame ledgers, joint-axis semantics, unit consistency, and spatial transforms before writing or regenerating `.urdf` artifacts. Core workflow: identify target consumers (RViz, robot_state_publisher, Gazebo, MoveIt, simulators), create or read design ledger, edit Python generator sources, regenerate explicit targets with `scripts/urdf`, validate on generation, and hand off results to cad-viewer. Generation-time validation fails fast on XML, graph, joint, geometry, mesh, and inertial problems. Integrates with CAD, SRDF, and cad-viewer skills for complete robot-description pipelines. Python gen_urdf() generator as source of truth; generated .urdf files treated as artifacts, not edited directly5kinstalls5SrdfThe srdf skill MoveIt2 SRDF generation, validation, and planning-semantics workflow. Use when creating, editing, regenerating, inspecting, or validating `.srdf` files, `gen_srdf()` sources, MoveIt planning groups, virtual joints, passive joints, end effectors, group states, disabled collisions, URDF-linked planning semantics, or SRDF handoff for live review. Use the URDF skill for robot structure, the SDF skill for simulator descriptions, and the cad-viewer skill for rendering, live review links, and optional MoveIt2 controls. Provenance: maintained in earthtojake/text-to-cad. Use the installed local skill files as the runtime source of truth; the repository link is only for provenance and release review. Use this skill for MoveIt semantic robot descriptions on top of an existing valid URDF. SRDF defines planning semantics; it does not define physical robot structure. SRDF correctness is a planning semantics problem. The common failure is not invalid XML; it is a plausible SRDF that gives MoveIt the wrong planning group, wrong tool link, wrong default state, unsafe disabled-collision matrix, or wrong joint units. Because language models are weak at spatial and4.9kinstalls6SendcutsendThe sendcutsend skill Review DXF and STEP/STP uploads for SendCutSend.com orders using its ordering guide, catalog, and specs. Use only for SendCutSend.com preflight reports covering upload readiness, selected material/SKU/thickness/service availability, and service-specific checks for laser cutting, CNC routing, bending, tapping, countersinking, hardware insertion, and finishing. Provenance: maintained in earthtojake/text-to-cad. Use the installed local skill files as the runtime source of truth; the repository link is only for provenance and release review. Use this skill to produce conservative, evidence-backed SendCutSend preflight reports for DXF and STEP/STP files. Treat SendCutSend's ordering guide, catalog JSON, and specs JSON as evidence feeds, not stable APIs. Field names, types, and coverage may vary. Do not turn missing, unparsable, N/A, or conflicting source data into a pass or fail. Fetch sources directly from official URLs and use local inspection code only to measure specific file facts; write the final report from explicit comparisons. Use the active project Python environment for local geometry inspection code. If the $cad4.9kinstalls7Cad ViewerCAD Viewer is a minimal agent skill interface for opening engineering files in a dedicated viewer from an agent session. The documented interface exposes display name CAD Viewer, a short description to open CAD, G-code, and robot files in Viewer, and a default prompt instructing the agent to start CAD Viewer and return review links for explicit CAD, G-code, or robot-description files named by the user. The skill does not describe conversion pipelines or modeling steps; it focuses on launching the viewer and surfacing review URLs once the user points at concrete file paths or filenames. Teams use it when an agent should hand off visual inspection of mechanical CAD exports, CNC G-code programs, or robot description assets without embedding a full CAD editor in chat. Invocation assumes the user supplies explicit file references rather than inferring paths. Output is review links suitable for quick human verification of geometry, toolpaths, or robot models before downstream manufacturing or simulation work.4.5kinstalls8Gcodegcode orchestrates plain FDM .gcode generation from 3D mesh files through real slicer CLI backends without uploading or starting print jobs. Supported inputs include .stl, .obj, unsliced .3mf, .ply, .glb, and .gltf, with .ply and glTF converted to temporary STL via optional trimesh at execution time. Every slice requires an explicit wrapper profile JSON pointing to absolute native slicer profiles plus machine and filament validation bounds. The workflow runs discover to pick orcaslicer, prusa-slicer, or curaengine backends, inspect for mesh readiness, dry-run slice commands, execute slice, then validate generated G-code for temperatures, movement, extrusion, and XYZ bounds. OrcaSlicer split profiles use native_settings and native_filaments arrays alongside native_config. After G-code work the skill mandates handing file paths to $cad-viewer when installed. Bambu-specific upload workflows are explicitly out of scope; validated plain .gcode hands off to $bambu-labs instead.3.9kinstalls9Bambu Labsbambu-labs is a hardware integration skill for local-network Bambu Lab print handoffs after validated plain .gcode exists. Safety rules default to dry-run plans, require --execute for real printer traffic, and mandate --execute --confirm-start-print before starting jobs while stating physical checks for build plate, filament, nozzle, and operator presence. Workflow steps validate G-code with $gcode, configure printers in workspace-root bambu-printers.json, read status, dry-run payloads, upload-only, then upload-start only after checks pass. Handoff modes include template-project for tested A1 Mini paths, plain for diagnostics, and bambox-project for supported P1S profiles. Print controls expose pause and cancel commands with dry-run defaults and explicit execute confirmations. After creating .gcode or .3mf artifacts, agents must hand file paths to $cad-viewer when installed. Developers reach for bambu-labs when agent-generated G-code should reach a Bambu printer over LAN FTPS and MQTT with cautious execute gates.3.8kinstalls10Implicit CadThe implicit-cad skill authors browser-native implicit CAD models as .implicit.js or .implicit.mjs modules exporting implicit.js/0.1.0 schema with GLSL sdf and optional color functions. It is experimental and prefers conventional STEP-first CAD unless the user explicitly wants implicit models. Modules may declare params becoming GLSL uniforms, animations with update callbacks, optional bounds and render settings, and procedural color via vec3 color. Authoring uses scripts/lib/implicit-cad.mjs helpers for primitives, smooth booleans, TPMS fields, and lattices without copying bundled files into outputs. Visual verification runs scripts/snapshot.mjs for single PNG or multi-camera job packets and optional animation GIFs. Export uses node scripts/export.mjs for GLB mesh handoff. After vault writes, handoff to $cad-viewer is mandatory when installed, returning live viewer links. Bounds auto-estimate from SDF unless animated or thin geometry needs explicit framing. Handoff to $cad-viewer is mandatory when that skill is installed.3.4kinstalls11DxfThe dxf skill Generate, regenerate, and validate 2D DXF drawings from Python ezdxf sources. Use for DXF files, gen_dxf sources, 2D profiles, outlines, templates, gaskets, panels, flat patterns, laser/plasma/waterjet cut layouts, and 2D drawing exports of CAD geometry. It covers standalone drafting : a Python source defining only gen_dxf . Use for pure 2D outputs - gaskets, panels, templates, cut layouts - with no 3D model behind them.. Key workflows include units: millimeters; set them explicitly on the document doc.units = ezdxf.units.MM .. 1. Convert the request into a short brief: outline dimensions, holes and slots, layers, units, output path, and validation targets. 2. For CAD projections, generate and validate the STEP geometry with $cad first, then add or update gen_dxf in the same source. When possible, derive the DXF from in-memory STEP/solid topology rather than duplicating geometry formulas, so the DXF remains a direct proj Developers invoke dxf when the task matches the triggers and reference files in SKILL.md for grounded, stepwise execution.3.2kinstalls12Cadcad is an agent skill from the text-to-cad library for creating, modifying, and validating parametric CAD models with STEP as the primary artifact. Reach for it when you need mechanical parts, assemblies, enclosures, or exportable mesh formats from plain-language specs, reference images, or existing STEP files. It pairs build123d source generation with inspection, mating, and fabrication-oriented outputs instead of treating CAD as one-off mesh dumps.0installs