Picogk

---
name: picogk
description: "Use when working on the silken_net Code-as-CAD surface — PicoGK voxel/SDF geometry generation in tools/cad/ (the Ti-coin coupon, the Zone-1 gyroid anode, the mechanical lock, the cathode flange, the PEEK radome) and the CEM (Computational Engineering Model) manifests that drive them; the voxel-FE elasticity tract (`fea`, `fea --fit` → Gibson-Ashby C/n pinned to canon by scripts/fea_canon_sync.rb); the resolution-adequacy and local-verify gates (`dotnet run -- verify`); the CEM provenance gate (every numeric field carries a declared ground); and the engineering-drawing tract (`draw <cem>` → SVG for people + DXF for the factory, an absent field prints NOT SPECIFIED IN CEM, the committed gallery under docs/images/cad lags its generator unless re-run). The non-obvious gotchas live in gotchas.md, one generated index line each in the body — open it before touching a generator, a manifest, draw, fea or verify; the task recipes live in tasks.md — open it before starting any CAD task. Routes to 01_02 §6 (stack, Noyron methodology, drawing norm) + 01_01 §5 (anchor geometry) + tools/cad/README; does not restate. Examples: \"generate the gyroid anchor\", \"add a CEM part / per-species SKU\", \"change anchor Ø / porosity / pore period\", \"why does the anchor crash with 'outside value 0'\", \"build the Ti-coin STL\", \"make a factory drawing / DXF\", \"add a surface-finish or coating note to a CEM\", \"what is the anchor lattice stiffness / run the FEA\", \"is this feature resolvable at this voxel\", \"set up the .NET CAD project\"."
---

# PicoGK Code-as-CAD (`tools/cad`)

Navigation aid + non-obvious gotchas. The **SSOT is the docs + code + `tools/cad/README.md`
below** — this skill points, it does not restate (so it can't drift). Verify a fact at its
home before trusting a summary. Methodology (per LEAP 71 Noyron): a CEM is a *deterministic
algorithm*, not generative ML — an agent writes the generator, the generator computes the geometry.

## SSOT Documents — Read These First

| Document | What it covers |
|----------|----------------|
| `tools/cad/README.md` | Operational home: layout, local-verify recipe, a short gotcha subset (the full list is this skill's `gotchas.md`), license |
| `docs/01_02_Ti_6Al_4V_Metallurgy_and_DMLS.md §6` | PicoGK stack (.NET 9, PicoGK 2.2 NuGet + ShapeKernel/LatticeLibrary submodules), Noyron methodology, real API, honest caveats — **and the engineering-drawing NORM**: why a drawing is derived from the CEM and not the mesh, the two readers (SVG clips, DXF does not), the loud-absence rule, what the acceptance contract must carry. The tools-local file below stays the RESEARCH + phase roster |
| `docs/01_01_Coaxial_Gyroid_Topology_and_PEEK.md §5/§6` | Anchor geometry: porosity nominal + acceptance band · pore gradient · Gibson-Ashby C/n and the MEASURED apparent stiffness per axis · Ti-coin Stage-2 coupon. ⛔ Values live THERE — this row names topics, never numbers |
| `docs/00_07_Action_Plan_Tracker.md` HW.1 / **HW.33** · HW.26 · HW.34 · HW.3 · HW.8 · HW.51 | Build state + the anchor geometry audit's OPEN legs (its ratified verdicts live in canon — `01_01 §5` · `02_02 §3–§4` · `01_04 §3.1/§5.5` · `02_01 §5.3`; the item's `**Стан:**` names each home). The CAD-facing open legs live in their own items: HW.26 the Zone-1 lock integration (G1 · G3 · G4) and the fate of the one groove left (its ring verdict is homed in `01_01 §4.3 B`) · HW.34 the pad-plane geometry and channel closure (after the HW.9 pin) and the coaxiality NUMBER (after HW.26 G1) — its ratified liner nominal and coaxiality FORM are homed in `01_01 §1.4` · HW.3 the Zone1↔2 fit band · HW.8 the placeholders (Zone-3 shank Ø, Zone-1 insertion) · HW.51 convergence (the radome numbers are ratified design points — `02_02 §3.5` · `02_02 §4.4`) |
| `docs/01_01_Coaxial_Gyroid_Topology_and_PEEK.md §6` | Cross-biome 5-SKU (pine/oak/broadleaf/mangrove/tropical) |
| `docs/00_03_TRL_Matrix_HIL_and_Beyond.md §3.6` | Code-as-CAD vs generative-AI; In-Silico for the Hardware stream |
| `tools/cad/docs/drawings_program.md` | Engineering-drawing program: CEM-native DXF (netDxf) + SVG, the ASME-Y14.5≠projection fix, lattice-as-inspection-card, phased §7 rollout |
| `extern/.../README_ImplicitLibrary.md` | LEAP's own implicit/TPMS guide — reference only; the graded v2 did NOT use it (own SDF — `tasks.md` «Graded gyroid») |

## Source Files

| File | Role |
|------|------|
| `tools/cad/cem/*.json` | CEM manifests — the Git-SSOT parameter inputs (`kind` discriminator: `ti_coin`, `anchor_zone1`, `mechanical_lock`, `cathode_flange`, `radome`, `zone2_sleeve`, `anchor_assembly`, `anchor_axial_stack`) |
| `tools/cad/src/SilkenCad/Program.cs` | CLI dispatch — ⛔ the verb roster is the `switch` in `Main`, never a list here. Outside it: `RunHeadless` (the `Library.Go` wrapper); `draw` is pure-managed (no Library.Go), `render`/`section` drive the native viewer |
| `tools/cad/src/SilkenCad/Cem.cs` | CEM records + JSON parse (snake_case). Engineering-drawing PMI lives here: optional `ToleranceSpec` (fits / interference µm / GD&T datums / a LIST of `Features` — a named feature with NO limits still renders, as loud absence) + `NotesSpec` (material/process/**post-process**/surface/coating-restriction/lattice-spec/inspection) on each part record — Noyron-native SSOT, fed to `draw` |
| `tools/cad/src/SilkenCad/Drawing.cs` | CEM-native engineering drawings (`draw <cem>`): **SVG (human) + DXF via netDxf (CAD-native factory deliverable, opens in AutoCAD/Fusion)** — no PDF (never built; `drawings_program §3`). Pure-managed string/entity build, no Library.Go. Consumes the CEM `ToleranceSpec`/`NotesSpec` (zero hard-coded eng-text); `DrawingStandard` param (ISO 1st-angle default / ASME). Each kind is a `Drawing.X` + `Drawing.XDxf` pair; the roster and the add-a-kind checklist → `tasks.md` «Generate an engineering drawing». ⚠️ gotcha #11 |
| `tools/cad/src/SilkenCad/TiCoin.cs` | Ti-coin coupon — `BaseCylinder` disc + `BaseRing` eyelet, `BoolAdd` |
| `tools/cad/src/SilkenCad/Zone1Anode.cs` | Zone-1 anode + `CartesianGyroid:IImplicit` (the from-scratch SDF) + `Anode()` render path |
| `tools/cad/src/SilkenCad/Validation.cs` | golden-metrics via `Voxels.CalculateProperties` (porosity needs an envelope ref) + reuses LEAP `Measure.fGetSurfaceArea` |
| `tools/cad/src/SilkenCad/Golden.cs` | Committed regression baseline of the GEOMETRIC metrics (`cem/<x>.golden.json`, `00_07` HW.33/HW.49) — written ONLY by `verify --write-golden` (the writer prints what moved), keyed by voxel (a run at another voxel is not compared). Pins REGRESSION, never correctness; ⚠️ a manifest with no golden passes `verify` silently |
| `tools/cad/src/SilkenCad/Connectivity.cs` | ARCH.25 two-phase topological audit — SDF-sample + 6-conn flood-fill (open/closed-pore, percolation, solid-island, specific-surface); pure-managed, display-less xUnit. ⊕ Also the MESH SOURCE for `VoxelFea`: `SampleRegion` (with a phase-override hook both FE callers leave null) and `LargestComponentMask` (membership, which `Components` does not keep) |
| `tools/cad/src/SilkenCad/TopologyCrossChecks.cs` | ARCH.25 third phase — Euler-χ / tortuosity / as-printed cross-checks, informational `verify` lines (gotcha #12); also the home of the lattice-thickness constants (`SheetThicknessPerPeriod` / `NetworkThicknessPerPeriod`, `LatticeThicknessMm`) that `Resolution` reads (gotcha #5) |
| `tools/cad/src/SilkenCad/WallScan.cs` | Verb `scan` — sweeps `wallParam` over the CEM's own SDF (pure-managed) and reports the working window (printable ∧ open-pore ∧ percolating); the sweep bounds follow the manifest's topology, because the param is a band on sheet and a level on network |
| `tools/cad/src/SilkenCad/VoxelFea.cs` | **Voxel-FE пружності** (00_07 HW.51/HW.33) — solid-воксель `Connectivity.Grid` → трилінійний гексаедр; матрично-вільний поелементний множник, 8-кольоровий за парністю вокселя, Якобі-CG. Верб `fea`. ⛔ Меш бере лише найбільше гране-звʼязне тіло (плавучий острівець = жорсткий рух), відкинуте — рапортується. Стеля оголошена в шапці класу; числа — `01_01 §5.2`, кеш `tools/cad/cache/fea/` |
| `tools/cad/src/SilkenCad/Resolution.cs` | **Достатність роздільності** (00_07 HW.51) — чи влазить фіча у воксель, якого просить її збірка. Ходить по РОЗПАРСЕНОМУ дереву записів, не по ключах json, і несе ВИВЕДЕНІ величини (щілина стрижень↔канал, товщина ґратки). HARD-носій — `ResolutionTests` + рядок у кожному `verify`-репортері |
| `tools/cad/src/SilkenCad/Probe.cs` | Verb `probe <voxel>` — falsifiable probes of what we believe about the PicoGK KERNEL, not our geometry (gotchas #1 · #9); probe B can kill the process by design, so A runs first |
| `tools/cad/src/SilkenCad/MechanicalLock.cs` | §4.3 mechanical lock — `MechanicalLockShank` asymmetric ratchet-barb SDF on the shank (Zone-1 solid shank, no channel / Zone-3 channelled); the DIN-471 groove is OPTIONAL per end (`MechanicalLock.HasGroove`, gotcha #0a); render = solid `voxConstruct` + thin barb-ridge `BoolAdd`; self-support metric (Noyron manufacturing-awareness) |
| `tools/cad/src/SilkenCad/CathodeFlange.cs` | Деталь 3 — Zone-3 cathode flange (Ø25): reuses the §4.3 shank/barbs via `ShankCem` + radial bayonet lugs + bus channel + the capsule's SINGLE face-seal O-ring groove on the top face — depth · width · radii DERIVED from the `o_ring` block and the socket band shared with the radome (`CathodeFlange.ORingGroove*`; 00_07 HW.33 branch (а), applied 2026-09-14), nothing stored; the flange sheet draws it as geometry in both readers and its depth tolerance as loud absence |
| `tools/cad/src/SilkenCad/Radome.cs` | Деталь 4 — PEEK radome v2c (Ø25): hollow dome + shield bell + a LOCAL INTERNAL RIM BOSS (`Radome.Boss*`/`SealLand*`: socket band `lug + clearance` outside ⊥ seal land `gland width + 2·clearance` inside; the rim-cavity ceiling (value: `02_02 §3.5`) — every term a MINIMUM, a ceiling HW.9 takes as an input, ⚖️ 2026-09-14) + the bayonet socket, cut in the boss's OUTER band only + PCB cavity; the rim is FLAT — the single O-ring groove is the flange's, derived from the `o_ring` block both manifests carry (gotcha #9 INVERTED — hollow is intended). ⛔ **ONE ratified-but-still-shipped divergence lives in this file:** the CAP is still a FULL HEMISPHERE where ⚖️ 2026-09-11 ratified a FLAT CROWN with an edge fillet (dims: `01_04 §5.5`), and it waits on ⚖️ HW.30 (piezo placement), not on a budget. ⚠️ Applying it silently inflates `hollow_fraction`, because `Validation`'s reference solid still adds a hemisphere — recompute that formula in the same commit — and `bell_rise_mm`/`bell_radius_mm` must become DRIVERS, which today they are not. ⚠️ The raised COLLAR that carries the lugs at `Assembly.RequiredLugZMm` is modelled on NEITHER part — its wall is set by nothing (no bayonet load model), and a placeholder would print on the flange sheet as a decision — so the socket keeps its L-slot shape (clipped to the outer band, pocket `socket band − skin` = 1.6 deep, skin 0.2) until that leg reshapes it: socket and collar are the female and male halves of one band. `verify radome` MEASURES the seal land solid over the full rim face (`SealLandSolidFraction`), the xUnit pins only derive it |
| `tools/cad/src/SilkenCad/Assembly.cs` | Capsule-end mate-audit (Деталь 3↔4, `02_02 §4.4`) — bayonet datum via `voxApplyTransformation` lift + Z/MATE-Ø/RF mismatch + skirt/inboard candidates; pure mate-math (xUnit) + render interference (`verify`) |
| `tools/cad/src/SilkenCad/Zone2Sleeve.cs` | Деталь 2 — Zone-2 PEEK sleeve (bore Ø11 / OD Ø15 / 50 mm): plain hollow tube via `BasePipe` (smooth bore; hex + flange-shoulder deferred, bench-gated) |
| `tools/cad/src/SilkenCad/AxialStack.cs` | Full axial stack mate-audit (Зони 1↔2↔3↔4, `02_02 §4.5`) — press-fit interference + insertion budget + span; reuses `Assembly.Build` + Zone-1 envelope; pure mate-math (xUnit) + render (`verify`) |
| `tools/cad/src/SilkenCad.Leap/` | vendored LEAP source compiled in (ImplicitUsings ON, warnings relaxed — not ours) |
| `tools/cad/extern/LEAP71_{ShapeKernel,LatticeLibrary}` | git submodules (source-only; not on NuGet) |

> Voxel/triangle counts, porosity AND the FE stiffness ratio all drift with voxel size — don't hardcode them ([[feedback_no_volatile_counts]]); MEASURE via `verify` (geometry) or `fea --sweep` / `fea --fit` (stiffness AND the Gibson-Ashby coefficients — a single row is an UPPER bound by construction, and the FITTED `C` itself still moves with the step (the `01_01 §5.2` table), so nothing is quoted without its convergence ladder).

**State** lives in `00_07` (the items in the table above), not here; memory keeps the router `[[project_picogk_code_as_cad]]` (⊕ its journal twin `[[log_picogk_cad]]` — the BODIES, with their numbers), and the curator lessons born in CAD sit in their class homes (applying a ratified verdict → `[[feedback_verdict_lifecycle]]`) · `[[project_anchor_bus]]` (bus router + the PEEK-liner trap, gotcha #0) · `[[project_coin_bakeoff_trl4]]` (the Ti-coin keystone this CAD feeds).

## Gotchas Not Obvious From Docs

**Bodies live in `gotchas.md` — open it before touching a generator, a CEM manifest, `draw`, `fea`
or `verify`.** Below is one generated line per gotcha: the line is the CARRIER, meant to stop you
mid-action; the mechanism and the bounds are in the companion. Numbering is
append-only (`0a` · `0b` · `4a` · `9b` are items of their own).

<!-- PICOGK-GOTCHAS-INDEX:AUTO — generated from gotchas.md by `ruby scripts/guard_craft_index.rb --write`; edit rules THERE, never here -->

0. `tools/cad/cem/*.json` is the parameter SSOT, and its `liner 0.15 mm` is now a RATIFIED branch — but read why before you touch it
0a. «CEM json = parameter SSOT» is true only where the manifest SPEAKS — and `cem_canon_sync` guards exactly that half, so the silent half has no gate at all
0b. A manifest can be silent for a SECOND reason, and this one is worse than a default: the record has no SLOT, so the key you write EVAPORATES
1. Render lattices via `new Voxels(IImplicit, BBox3)` + `BoolIntersect`
2. `ImplicitRadialGyroid` is degenerate near r=0 — use a cartesian gyroid for a small part near its own axis
3. Run headless through `Library.Go(voxel, task, bEndAppWithTask:true)`, never `new Library()` — and headless CI cannot run it at all
4. Gyroid `wallParam` is a DIMENSIONLESS level of the gyroid equation, and the sheet and network branches read it differently
4a. «Porosity is voxel-dependent → MEASURE it», and the suspected cause of the coarse-voxel collapse is REFUTED — both now live in #4 (evidence: `00_07` HW.49)
5. Voxel-resolution floor + gradient distortion — the grid's floor and the printer's floor are different ceilings, and the printable one depends on topology
6. Keep `ImplicitUsings` ENABLED for the vendored `src/SilkenCad.Leap`, and the strict knobs ON only for our own code
7. Set `DOTNET_ROOT=$HOME/.dotnet` when running the apphost binary directly, and look for its lowercase name `silkencad`
8. Connectivity needs WALL-resolution, not pore-resolution — sample the SDF finer than half the thinnest wall, and judge by the COUNT
9. A FILLED (solid) body must come from ShapeKernel `voxConstruct`, NOT `new Voxels(IImplicit, BBox3)` — unless your field is CLOSED, and that exception is measured, not theoretical
9b. A body can render PERFECTLY in its own grid and add NOTHING to the assembly — and its own volume is the metric that hides it
10. `render`/`section` (presentation) need a display + the screenshot is TGA
11. A drawing must never invent or silently drop a field — an absent value prints `NOT SPECIFIED IN CEM`, and the next `??`-default will read as innocent as the four fixed on 2026-08-28
12. `TopologyCrossChecks.cs` (ARCH.25 third phase, 2026-09-09) — Euler-χ / tortuosity / as-printed cross-checks, wired into `verify` as `[nice-to-have]` INFORMATIONAL lines, not the `VERIFY OK/FAILED` gate
13. Every numeric field of every `cem/*.json` must carry a DECLARED GROUND, and this is a HARD gate since 2026-09-12 — adding a field without one reds `docs.yml`
14. A bare `fea --fit` OVERWRITES a canon-pinned cache with a DIFFERENT SPECIMEN, because the filename carries the RESOLUTION and not the SIZE
15. A direction or a sign in a CEM field is read in TWO frames — the author's assembled/world view and the code's part-local frame — so pin the FUNCTION in the part's frame, never the sign
16. A constant-period lattice sampled at a step that divides its period is PHASE-LOCKED — and `--step-div N` takes the step FROM the period, so on a constant-period part EVERY divisor is such a step, by definition
17. `out/` is OUTSIDE every gate we own, and that is where the VENDOR's copy lives
18. A mutation probe that restores a `.cs` with an OLDER mtime is not restored for MSBuild — and the next probe's red is the previous probe's leftover

<!-- /PICOGK-GOTCHAS-INDEX -->

## Common Tasks

**Recipes live in [`tasks.md`](tasks.md) — open it before you START any CAD task** (a part or SKU, the anchor geometry, the bus wire, a lattice, a drawing, a render, an FE or resolution run, `verify` or `converge`, an assembly audit, a submodule bump). They carry live verdicts and traps inline, so never work a task from memory of this skill.