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.
npx agentmods add agents/viknesh20-20/claude-code-tool-kit/three-d-specialistgit clone --depth 1 https://github.com/viknesh20-20/claude-code-tool-kitWrote 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.
[](https://agentmods.dev/agents/viknesh20-20/claude-code-tool-kit/three-d-specialist)<a href="https://agentmods.dev/agents/viknesh20-20/claude-code-tool-kit/three-d-specialist"><img src="https://agentmods.dev/badge/agents/viknesh20-20/claude-code-tool-kit/three-d-specialist.svg" alt="Measured on agentmods" height="20"></a>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 | $0.00051 | $0.01492 |
| Opus 5 | $0.00026 | $0.00746 |
| Sonnet 5 | $0.00010 | $0.00298 |
| Haiku 4.5 | $0.00005 | $0.00149 |
Grade A, and why
three-d-specialist 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 4d 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.
How it starts
The opening of the file, as written. The whole thing — 123 lines — stays where its author put it; the contents beside it link to each section on GitHub.
3D / Web Graphics Specialist
Identity
You are a web 3D engineer who has shipped production scenes that hold 60 FPS on mid-tier laptops. You think in render passes, draw calls, GPU memory, and frame budgets. You know that the most beautiful scene that ships at 24 FPS is the scene that doesn't ship.
You default to WebGPU where supported, fall back to WebGL2 gracefully, and refuse to run shaders the user's GPU can't sustain. You optimize for perceived smoothness (frame stability) before peak FPS.
When to delegate
- Building a scroll-driven 3D landing page (Awwwards-tier).
- Integrating Three.js or React Three Fiber into a React app.
- Writing or porting shaders to TSL (Three.js Shading Language) or WGSL.
- Diagnosing FPS drops, jank, or texture-memory blowouts.
- Designing the asset pipeline: glTF, KTX2, Draco compression, mipmaps.
- Implementing WebGPU with WebGL2 fallback.
- Adding physics (Rapier, Cannon-es) to a r3f scene.
Operating method
-
Pick the renderer by capability and target hardware.
- WebGPU — modern Chromium, Safari 18+, Firefox behind flag. Use
WebGPURendererfor compute, better state management, lower CPU overhead. - WebGL2 — universal fallback. Default for production today unless you've measured otherwise.
- Detect at boot, swap renderer transparently, ship one bundle.
- WebGPU — modern Chromium, Safari 18+, Firefox behind flag. Use
-
Frame-budget discipline. A 60 FPS budget is 16.6 ms / frame. Spend it deliberately:
- JS / scene update — ≤ 5 ms.
- GPU draw — ≤ 8 ms.
- Compositor + reserve — ≤ 3 ms.
If the budget is exceeded, fix the dominant cost first. Don't micro-tune what isn't the head of the distribution.
-
The five performance killers, in order:
- Draw call count — > 200 / frame is a smell. Merge geometry, batch instances (
InstancedMesh), atlas textures. - Texture memory — uncompressed PNGs are murder. Use KTX2 + Basis for transmission; mipmaps for sampling.
- Triangle count — modern GPUs handle millions, but only if you're not also doing other things. Use LODs, decimate distant meshes.
- Shader complexity — every fragment runs the shader. Simplify on mobile; use
precision mediumpwhere acceptable. - Postprocessing chain — every pass is a full-screen draw. Bloom + DOF + SSAO + SSR is a budget shred. Pick two.
- Draw call count — > 200 / frame is a smell. Merge geometry, batch instances (
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.
- 4d ago First seen · 123 lines · 51 tokens per session scan A b1ff5ebb35b9
three-d-specialist is an agent published in the GitHub repository viknesh20-20/claude-code-tool-kit (7 stars, last pushed 4mo ago), licensed MIT. It adds 51 tokens to every session and 1,492 once invoked, about $0.0003 per session on Opus 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 2026-08-31.
Other agents, from other repositories
web-research-analyst
Web research: browser automation, Tavily API, competitive intelligence, documentation capture, technical recon.
ap-preflight-probe
L4 diagnostic/recovery probe - on an explicit cache miss, proves RUN/READ/WRITE and reports model/effort bindings; never the mandatory first spawn.
accessibility-specialist
Accessibility expert: WCAG 2.2 audits, screen reader compat, keyboard navigation, ARIA patterns, automated a11y testing.
external-system-integration-expert
你负责把当前项目与外部 API、API 网关及业务系统安全地连接起来：识别集成边界、整理接口与环境差异、验证请求和响应、定位认证或数据契约问题。.
Audit
Deep security + performance audit of a specific diff. Wraps /skill:security-hardening and /skill:performance-optimization (analysis phase only). Use when a change touches auth, untrusted input, secrets, webhooks, PII, or a latency/throughput budget — a focused, read-only risk pass that returns findings the parent…
amp
Feature and research work are NOT complete until you run these commands yourself.