Agent skill

Debugging Geometry

by andymai in andymai/brepjs

This skill should be used when debugging a geometry failure in brepjs — when a task says "boolean returned an invalid shape", "fuse/cut/intersect returned Err", "the result is empty", "measureVolume…

Apache-2.0Auto-check passedDevelopment

Install Debugging Geometry

skills CLI
$ npx skills add andymai/brepjs --skill debugging-geometry -a claude-code

Project install by default; add -g for ~/.claude/skills/.

GitHub CLI
$ gh skill install andymai/brepjs debugging-geometry --agent claude-code

Project scope by default; add --scope user for a personal install. Needs GitHub CLI 2.90.0 or later (public preview).

Manual copy
$ git clone --depth 1 https://github.com/andymai/brepjs.git skills-src && mkdir -p .claude/skills && cp -r skills-src/.claude/skills/debugging-geometry .claude/skills/debugging-geometry && rm -rf skills-src

Use ~/.claude/skills/ instead of .claude/skills for a personal install. The folder must contain SKILL.md.

Claude Code skills documentation · loads skills from .claude/skills/

Facts

Skill name
debugging-geometry
GitHub stars
114
Token cost
~3.8k tokens
SKILL.md length
1,522 words
Files
4 (incl. references)
Skills in repo
21
Repo updated
First seen
Licence
Apache-2.0

At a glance

This skill should be used when debugging a geometry failure in brepjs — when a task says "boolean returned an invalid shape", "fuse/cut/intersect returned Err", "the result is empty", "measureVolume…

  • Works in 7 steps: Triage by failure class → Numeric sanity before trusting any render → Invalid shapes and the healing pipeline → …
  • Tasks that involve Debugging
  • SKILL.md covers Step 1 — Triage by failure class, Step 2 — Numeric sanity before…, Step 3 — Invalid shapes and… and Step 4 — Failed booleans, plus 5 more sections
  • Calls npm

What it does

Debugging Geometry is an agent skill from andymai/brepjs. This skill should be used when debugging a geometry failure in brepjs — when a task says "boolean returned an invalid shape", "fuse/cut/intersect returned Err", "the result is empty", "measureVolume returns 0", "autoHeal didn't fix it", "shape is invalid", "STEP export crashed" or "WebAssembly.RuntimeError during export", "the part looks wrong, render it", "works on occt-wasm but fails on brepkit", or "is this a kernel divergence". Covers triage, numeric sanity checks, the healing pipeline, boolean failure modes…

Its SKILL.md is about 3.8k tokens, which your agent loads only when the skill is triggered. The skill folder holds 4 other files, including reference files (for example `references/boolean-failures.md`, `references/healing.md` and `references/kernel-divergence.md`).

It sits in Development, covering Debugging. It works with WebAssembly. The repository describes itself as: Web CAD library with exact B-Rep geometry. The licence is Apache-2.0.

When your agent uses it

  • Tasks that involve Debugging

Example prompts

  • “boolean returned an invalid shape”
  • “fuse/cut/intersect returned Err”
  • “the result is empty”
  • “/debugging-geometry”

Workflow steps

7 steps, taken from the step headings in SKILL.md.

  1. Triage by failure class
  2. Numeric sanity before trusting any render
  3. Invalid shapes and the healing pipeline
  4. Failed booleans
  5. Visual debugging with the brep CLI
  6. Kernel-divergence isolation
  7. Writing the repro test

What it can do on your machine

Read from SKILL.md and the folder at commit e8fa511. It shows what the files ask for, not the result of running them.

  • Tool permissions

    Pre-approves nothing: there is no allowed-tools line, so your agent's usual permission prompts apply.

    From allowed-tools in the SKILL.md frontmatter.

  • Runs code

    Shell commands in SKILL.md call:

    • npm

    From the folder's file list and the shell code blocks in SKILL.md.

  • Network

    No URLs in SKILL.md. Its commands use npm, which can reach the network depending on how they are called.

    From URLs in SKILL.md, links to its own repository left out.

  • Credentials

    Names no API keys, tokens, secrets or passwords.

    From names ending in _API_KEY, _TOKEN, _SECRET, _KEY or _PASSWORD in SKILL.md.

Context cost

Debugging Geometry loads about 3.8k tokens when it runs, and up to ~6.6k if it reads all its reference files. Until then it costs about 138 tokens; SKILL.md has 1,522 words of instructions outside code blocks.

Always · name and description, kept in context so the agent knows when to use it
~138
When it runs · the whole SKILL.md, loaded when a task matches
~3.8k
With references · SKILL.md plus every file in references/, read only if the agent opens them
~6.6k

Estimates: characters ÷ 4, the usual rule of thumb; real counts depend on the model's tokenizer. Scripts and assets cost tokens only if the agent reads them.

Safety

Auto-check passed

The automated check found no risky patterns in SKILL.md.

Automated static check — not a guarantee. Review scripts before installing. It scans the text of SKILL.md for risky patterns (piping downloads into a shell, reading credential files, hidden Unicode, destructive commands); files beside SKILL.md are not scanned.

SKILL.md

The full file from andymai/brepjs at commit e8fa511, republished under its Apache-2.0 licence (© andymai). 1,522 words, ~3,816 tokens.

Download SKILL.mdSave it as .claude/skills/debugging-geometry/SKILL.md (or your agent's skills folder). This skill also uses 3 other files; get the full folder from GitHub.
name
debugging-geometry
description
This skill should be used when debugging a geometry failure in brepjs — when a task says "boolean returned an invalid shape", "fuse/cut/intersect returned Err", "the result is empty", "measureVolume returns 0", "autoHeal didn't fix it", "shape is invalid", "STEP export crashed" or "WebAssembly.RuntimeError during export", "the part looks wrong, render it", "works on occt-wasm but fails on brepkit", or "is this a kernel divergence". Covers triage, numeric sanity checks, the healing pipeline, boolean failure modes (including the

Debugging geometry failures

A symptom-ordered playbook for the five failure classes: an Err result, wrong-but-Ok geometry, an empty/degenerate result, a crash/trap, or a kernel divergence. Verify every hypothesis with numbers and a render before changing code.

Step 1 — Triage by failure class

SymptomClassFirst move
Op returned Result.Errerror resultRead error.code + error.suggestion + error.metadata?.diagnostics
Result is Ok but geometry is wrongbad geometryNumeric sanity (Step 2), then render (Step 5)
Result is empty / measureVolume==0degenerate/emptyisEmpty, getBounds, volume-normalization rule (Step 2)
Session crashed / WebAssembly.RuntimeError on exportkernel trapThe #1126 export-crash class (Step 4 + references/boolean-failures.md)
Passes on one kernel, fails on anotherdivergenceIsolate under TEST_KERNEL (Step 6)

Every fallible op returns Result<T, BrepError> with kind/code/message/suggestion?/cause?/metadata?. The full code catalog with per-code recovery advice lives in docs/errors.md — look codes up there, don't guess. error.suggestion is a first-class field; read it first (src/core/errors.ts).

Step 2 — Numeric sanity before trusting any render

Confirm the defect with numbers before opening an image; a render can hide a zero-thickness sliver or an open shell.

  • measureVolume / measureArea / measureLength / measureDistance in src/measurement/measureFns.ts — all return Result<T> and pre-check null → NULL_SHAPE_INPUT.
  • volume === 0 is itself a diagnostic. Volume is normalized to 0 for anything that is not a solid/compsolid/compound (measureVolumeProps, measureFns.ts, kernel divergence #1361). A "solid" reporting 0 volume is almost always an open shell that never got sewn/closed. Check describe(shape) and the shape type before assuming the volume math is wrong.
  • isEmpty(shape) (src/topology/shapeFns.ts) is only a kernel null check — it does not detect degenerate-but-non-null geometry.
  • getBounds(shape) (src/topology/topologyQueryFns.ts) returns the AABB and can throw on degenerate shapes — that throw is the basis of the pre-export probe (Step 4).
  • describe(shape) (src/topology/topologyQueryFns.ts) gives a structural fingerprint (type + sub-shape counts) — use it to confirm "is this actually a solid, or a compound of two solids?".
  • Measurements and validity are cached per shape. After healing or any in-place kernel repair, a stale isValid/measurement can lie — call invalidateShapeCache (exported from the index) if a value looks stale.

Step 3 — Invalid shapes and the healing pipeline

Full tables in references/healing.md. Core loop:

  1. Check isValid(shape) (src/topology/healingFns.ts).
  2. autoHeal(shape, options?) → Result<{ shape, report: HealingReport }>.
  3. Escalate only if report.isValid is still false.

Critical caveat — the short-circuit. autoHeal returns immediately when the input is already valid: report.alreadyValid === true, steps === ['Shape already valid'], and diagnostics contains ONLY { name:'validation', attempted:true, succeeded:true } (healingFns.ts). The absence of sew/fixSelfIntersection/healSolid diagnostics does not mean those passes found nothing — they never ran. Never conclude "healing found no problems" from an alreadyValid report.

Two more traps:

  • wiresHealed/facesHealed are Math.abs(after - before) count deltas, a heuristic change-detector — not a count of repaired defects.
  • Sewing only runs when sewTolerance is passed; fixSelfIntersection defaults to false.

Escalation ladder when autoHeal leaves it invalid: pass a sewTolerance → fixShape (general ShapeFix_Shape) → solidFromShell for a shell that should be a solid → give up with HEAL_SOLID_INCOMPLETE / HEAL_NO_EFFECT. All in healingFns.ts; details in references/healing.md.

Step 4 — Failed booleans

Full recipes in references/boolean-failures.md. Two distinct diagnostic surfaces — do not conflate them:

  • checkBoolean(base, tool, op) (src/topology/booleanDiagnosticFns.ts) is a pre-flight predictor: it returns { valid, issues } where issues are only 'null-shape' | 'not-valid' per operand. It predicts failure; it does not explain a failure after the fact.
  • BooleanDiagnostics ({ hasErrors, hasWarnings, messages }) rides on results/errors when trackEvolution is on (default). messages is currently always empty — OCCT's Standard_OStream reporting is not reachable in WASM builds (src/kernel/types.ts). Rely on hasErrors/hasWarnings, not on message text. On hasErrors + null result the op retries without evolution tracking; on hasErrors + non-null it warns and continues (booleanFns.ts).

Boolean ops pre-validate null operands → NULL_SHAPE_INPUT before touching the kernel (booleanFns.ts). When the result cannot cast to 3D the error names the actual type ("Got COMPOUND instead.", booleanFns.ts) — a strong signal the boolean degenerated.

Symptom → cause → fix:

SymptomLikely causeFix
FUSE_* Err, "Got COMPOUND"operands don't actually overlapgive a real overlap; for a weld, fuseAll(shapes, { unsafe: true })
BOOLEAN_HAS_ERRORScoincident/near-tangent facesperturb one operand slightly, or set fuzzyValue
overlapping coplanar / zero-thickness inputnon-manifold inputautoHeal() operands first (the baked-in FUSE suggestion)
in-memory checks all pass, STEP writer traps#1126 disjoint-fuse corruptionfuseAll(shapes, { strategy: 'pairwise' })

The #1126 export-crash class is the nastiest: fuseAll with the default strategy: 'native' (N-way BRepAlgoAPI_BuilderAlgo) can silently corrupt certain disjoint inputs so the result passes isValid, validSolid, mesh, measureArea, and getBounds — yet traps the STEP writer with a WebAssembly.RuntimeError that corrupts the Emscripten heap and poisons the kernel for the rest of the session (meshFns.ts). No known non-trapping check detects it; the only safety nets are the pre-export bounding-box probe (probeSerializable, catches some degenerates → STEP_EXPORT_UNSERIALIZABLE/STL_EXPORT_UNSERIALIZABLE with the offending sub-solid localized) and exportError classifying the trap as *_EXPORT_CRASHED. The fix is strategy: 'pairwise' (divide-and-conquer over BRepAlgoAPI_Fuse, a different algorithm that is unaffected; booleanFns.ts). Tracked upstream at andymai/opencascade.js#3.

Where export lives (structural gotcha). Import functions live in src/io/ (importFns.ts), but the STEP/STL/IGES export functions live in src/topology/meshFns.ts (exportSTEP/exportSTL/exportIGES) and assembly STEP in src/operations/exporterFns.ts — not in io/. importSTL auto-runs ShapeUpgrade_UnifySameDomain on the read shape (src/kernel/occt/ioOps.ts); IGES round-trips both ways.

Finder mis-selection (fillet/chamfer hit the wrong edges)

When a modifier lands on the wrong entities, the defect is the selection, not the op. Finder USAGE (predicates, when/inList/not, sorting) lives in the docs finders page (apps/docs/tasks/finders.md) and src/query/README.md; this skill owns the debugging angle. Core surface: findAll(shape) returns every match, findUnique(shape) returns Result<T> and errors when 0 or >1 match (src/query/finderCore.ts, findUniqueIn).

SymptomLikely causeFix
Fillet/chamfer landed on extra/other edgespredicate too loose — more entities matched than intendedfindAll to count what actually matched, render with brep snapshot --label to see them, then tighten the predicate
findUnique returned Err (0 or >1 match)predicate too strict (0) or ambiguous (>1)Inspect findAll length; add a discriminating clause or switch to findAll if many are intended
normalAt(face) / a normal comes back null on the picked entitythe selected face/edge is degenerate, so the surface normal is undefined (normalAt, src/topology/faceFns.ts)Confirm with describe/getBounds (Step 2); the finder selected a degenerate sub-shape — re-find geometrically or heal first
Show full SKILL.md (544 more words)Show less

Step 5 — Visual debugging with the brep CLI

The brep CLI (bin in packages/brepjs-cad) renders and validates .brep.ts modules. For interpreting the verify report (checks, hints, body relations, the multi-body fragmentation advisory), defer to the sibling skill brepjs:verify (packages/brepjs-cad/skills/verify/SKILL.md) — this skill covers the debugging loop, not the report anatomy.

  • brep verify <file>.brep.ts [--check] [--json] [--metrics] is the primary loop. --check typechecks first; --metrics adds deterministic manufacturability metrics + the interference matrix (off by default, slow). --expect-code <CODE> / --expect-invalid assert a known-bad part fails the right way.
  • brep snapshot <file>.brep.ts [--out dir] [--label tag] renders all views in one pass: iso, front, top, right, iso-xray (reveals bores/internal walls an opaque render is blind to), plus a conditional aimed section shot and a marks (Set-of-Marks B#/H# label) shot when the report detects bores/bodies (snapshot/shoot.ts). Each PNG has its bbox size (W×D×H) burned in, so scale is readable from the image. --label writes to a subfolder for pre/post A/B pairing.
  • Snapshots use a private ephemeral render server (fresh: true), so parallel snapshots don't contend on the shared :7373 server.
  • Puppeteer is lazily/optionally imported: without it, the JSON report still emits and the CLI prints snapshots need puppeteer/Chrome — run: npm i puppeteer.
  • Prefer one brep snapshot over rendering each angle separately — one pass gives every standard view. Judge fine surface finish (fine threads, small chamfers) from a real playground render rather than the snapshot, which can exaggerate thin features.

Compare geometry with brep diff <a> <b> (volumeDelta/areaDelta/bboxDelta/symmetricDifferenceVolume) and brep measure <a> [b].

Step 6 — Kernel-divergence isolation

Full mechanics in references/kernel-divergence.md. When a part passes on the default kernel but fails elsewhere (or vice-versa):

  1. Re-run under the other kernel. npm run test:occt and npm run test:brepkit set TEST_KERNEL and run that vitest project (vitest.config.ts; tests/setup-kernel.ts). CI runs only occt-wasm (.github/workflows/ci.yml), so brepkit/manifold divergences must be reproduced locally.
  2. Localize the diverging op. The in-repo technique (tests/kernelDivergenceCoverage.test.ts): compare against an analytic reference AND against an alternate representation of the same shape — e.g. a torus primitive is exact while a revolve sweep undershoots its volume by ~2%, which pins the loss to the sweep, not the primitive (#968).
  3. When confirmed, register it in tests/helpers/kernelDivergences.ts (single source of truth; key = operation.specificCase, kinds not-implemented | skip | tolerance | topology-differs) and gate tests with skipIfDiverges(ctx, key). Always cite the upstream issue. The conformance matrix docs/kernel-conformance.md is auto-generated (npm run conformance:generate).

Note: the default occt-wasm kernel exposes no raw oc handle (kernelDivergences.ts) — raw-OCCT debugging tricks (poking TopoDS_*, patching FS) are unavailable there.

Step 7 — Writing the repro test

  1. import { initKernel } from './setup.js' (initOC is a legacy alias) and beforeAll(async () => { await initKernel(); }, 30000).
  2. Use unwrap(result) in tests; assert geometry with toBeCloseTo(expected, precision) — never exact float equality.
  3. To force the invalid-shape healing path deterministically, spy kernel.isValid to return false on the first call then delegate — the mockKernelIsValid pattern in tests/autoHeal.test.ts. See the sibling skill writing-tests.

Additional resources

  • references/healing.md — HealingReport/diagnostic-name tables, AutoHealOptions, escalation ladder, short-circuit semantics.
  • references/boolean-failures.md — checkBoolean vs BooleanDiagnostics, error codes + baked-in suggestions, fuzzyValue, the #1126 case study.
  • references/kernel-divergence.md — registry mechanics, divergence kinds, kernel capability table, occt-wasm no-raw-oc limits, brepkit #965–968 family status.
  • docs/errors.md — full error-code catalog. docs/getting-started.md §Troubleshooting — the 4-step boolean recovery recipe. docs/kernel-conformance.md — capability matrix.
  • Sibling skills: result-error-handling (Result/BrepError construction), writing-tests (test skeleton + divergence skips), kernel-abstraction (adding kernel methods), brepjs:verify (CLI report interpretation).

© andymai, Apache-2.0. Rendered from Markdown: HTML in the file is shown as text, images as links, and headings moved down two levels. Raw file

Files

SKILL.md and 3 other files (references) in .claude/skills/debugging-geometry of andymai/brepjs.

  • SKILL.md
  • references/boolean-failures.md
  • references/healing.md
  • references/kernel-divergence.md

Open the folder on GitHubat commit e8fa511

Compare with similar skills

Debugging Geometry next to the 5 skills that share the most tags, products or categories with it. Stars are the repository's; “used in” counts other GitHub owners with a copy.

Debugging Geometry compared with similar skills
SkillStarsUsed inTokensAuto-checkLicenceRepo updated
Debugging Geometry this skillandymai/brepjs114—~3.8kAutomated safety check: PassApache-2.0
Ssh Connection ImportfeigeCode/navop1.8k—~1.2kAutomated safety check: WarnCustom licence
Optimizer Debugwado-lang/wado116—~2.5kAutomated safety check: PassMIT
Chrome Devtools MCPmanagedcode/dotnet-skills486—~2.2kAutomated safety check: PassMIT
Wasm Emscriptenmohitmishra786/low-level-dev-skills253—~1.7kAutomated safety check: PassMIT
Create Featurezacharyfmarion/openscad-studio238—~1.5kAutomated safety check: PassGPL-2.0

Similar skills

  • Ssh Connection Import

    feigeCode/navop

    A skill your agent uses when implementing or debugging onetcli SSH connection importers, including OpenSSH config, knownhosts, candidate file manifests, local composite installs, SSH authmethod JSON…

    1.8k GitHub stars~1.2k tokensUpdated today
    DevelopmentAuto-check: warnings
  • Optimizer Debug

    wado-lang/wado

    Debug the Wado optimizer with the WADOTRACE, WADODUMPPASSBEFORE/AFTER, WADOLISTPASSES, and WADOSKIPPASS env vars.

    116 GitHub stars~2.5k tokensUpdated today
    DevelopmentAuto-check passed
  • Chrome Devtools MCP

    managedcode/dotnet-skills

    Use Chrome DevTools MCP from .NET agents and .NET-focused repos to inspect, debug, and automate Chrome through an MCP client.

    486 GitHub stars~2.2k tokensUpdated today
    DevelopmentAuto-check passed
  • Wasm Emscripten

    mohitmishra786/low-level-dev-skills

    WebAssembly with Emscripten skill for C/C++ to WASM compilation.

    253 GitHub stars~1.7k tokensUpdated 3 mo ago
    DevelopmentAuto-check passed
  • Create Feature

    zacharyfmarion/openscad-studio

    A skill your agent uses when the user asks to take a feature or bug fix from prompt to implementation, especially prompts like "/create a new feature", "build this feature end-to-end", "take this…

    238 GitHub stars~1.5k tokensUpdated 1 mo ago
    Agent WorkflowsAuto-check passed
  • Flowfile Node Development

    Edwardvaneechoud/Flowfile

    End-to-end runbook for adding or modifying a Flowfile node type across all four layers (flowfilecore settings/graph/template, flowfilefrontend UI registry, flowfileframe Python API, flowfilewasm…

    370 GitHub stars~9.3k tokensUpdated today
    DevOps & CloudAuto-check passed

More from andymai/brepjs

All 21 skills in this repo
  • Implement

    andymai/brepjs

    A skill your agent uses when authoring or editing a brepjs .brep.ts part — writing the geometry with the functional API (box, cylinder, fuse, cut, fillet, sketch→extrude…), declaring an expected…

    114 GitHub stars~3.4k tokensUpdated today
    Auto-check passed
  • Memory And Disposal

    andymai/brepjs

    This skill should be used when managing WASM handle lifetimes or hunting memory leaks in brepjs — when a task mentions "createHandle() without using keyword risks WASM memory leak"…

    114 GitHub stars~3.1k tokensUpdated today
    Auto-check passed
  • Polish

    andymai/brepjs

    A skill your agent uses when a valid brepjs part should look designed rather than glued-from-primitives (products, toys, mechanisms, anything a human eyeballs), and when exporting/handing off the…

    114 GitHub stars~588 tokensUpdated today
    Auto-check passed
  • Wasm Interop

    andymai/brepjs

    This skill should be used when working across the JS/WASM boundary in brepjs — writing or debugging code in src/kernel/occt, src/kernel/occtWasm, or src/kernel/brepkit, or diagnosing symptoms like…

    114 GitHub stars~3k tokensUpdated today
    Auto-check passed
  • Writing Tests

    andymai/brepjs

    This skill should be used when writing, running, or fixing tests in the brepjs repository — when a task says "add a test", "write a regression test", "tests are failing", "test timed out", "coverage…

    114 GitHub stars~4.3k tokensUpdated today
    Auto-check passed
  • Adding Operations

    andymai/brepjs

    This skill should be used when adding or extending a geometric shape operation in brepjs — the end-to-end recipe once the target module is chosen (which is decided by architecture-navigation) — when…

    114 GitHub stars~4.4k tokensUpdated today
    Auto-check passed

Works with

Categories

Questions about Debugging Geometry

What does Debugging Geometry do?

This skill should be used when debugging a geometry failure in brepjs — when a task says "boolean returned an invalid shape", "fuse/cut/intersect returned Err", "the result is empty", "measureVolume…. Debugging Geometry is an agent skill from andymai/brepjs.RuntimeError during export", "the part looks wrong, render it", "works on occt-wasm but fails on brepkit", or "is this a kernel divergence".

When should I use Debugging Geometry?

Debugging Geometry fits situations like: tasks that involve Debugging.

How do I install Debugging Geometry in Claude Code?

Run `npx skills add andymai/brepjs --skill debugging-geometry -a claude-code`. Or copy the skill folder (.claude/skills/debugging-geometry in andymai/brepjs) into .claude/skills/debugging-geometry in your project. Claude Code loads it when a task matches its description.

How do I install Debugging Geometry in Codex?

Run `npx skills add andymai/brepjs --skill debugging-geometry -a codex`. Or copy the skill folder (.claude/skills/debugging-geometry in andymai/brepjs) into .agents/skills/debugging-geometry in your project. Codex loads it when a task matches its description.

Can I use Debugging Geometry in Cursor, Gemini CLI or GitHub Copilot?

Cursor, Gemini CLI, GitHub Copilot and OpenCode also load SKILL.md folders. With the skills CLI, run `npx skills add andymai/brepjs --skill debugging-geometry -a cursor` (or -a gemini-cli, github-copilot or opencode for the others). To copy it by hand, put the folder in .cursor/skills/debugging-geometry, .gemini/skills/debugging-geometry, .github/skills/debugging-geometry and .opencode/skills/debugging-geometry in your project.

What does Debugging Geometry need to run?

Going by SKILL.md and its folder, Debugging Geometry needs the command-line tools its instructions call (npm).

Does Debugging Geometry access the network?

SKILL.md contains no URLs. Its commands use npm, which can reach the network depending on how they are called. This is read from the text; nothing was executed.

Is Debugging Geometry safe to install?

Our automated static check of SKILL.md found no risky patterns, such as piping downloads into a shell, reading credential files or hidden Unicode. It is not a guarantee. Review the folder before installing.

What licence does Debugging Geometry use?

Debugging Geometry is published under the Apache-2.0 licence (the repository's licence). It allows redistribution, so the full SKILL.md is shown on this page.

How many tokens does Debugging Geometry use?

About 3.8k tokens (SKILL.md is roughly 15k characters). Agents keep only the skill's name and description in context until a task matches; then they load SKILL.md in full. Its references folder adds about 2.8k tokens, read only when the agent opens those files.

What are the alternatives to Debugging Geometry?

Skills that share tags, products or a category with Debugging Geometry: Ssh Connection Import (feigeCode/navop, 1.8k stars), Optimizer Debug (wado-lang/wado, 116 stars), Chrome Devtools MCP (managedcode/dotnet-skills, 486 stars) and Wasm Emscripten (mohitmishra786/low-level-dev-skills, 253 stars). The comparison table on this page puts their stars, adoption, token cost, safety result and licence side by side.

Who maintains Debugging Geometry?

andymai (a GitHub user) maintains it in andymai/brepjs, which has 114 GitHub stars. The repository holds 21 skills in this directory. The repository was last updated on October 6, 2026.

Source: andymai/brepjs on GitHub. Facts on this page come from the repository at the commit we read; the author's words are quoted as theirs.