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 skills add cl0nazepamm/3dsmax-mcp --skill 3dsmax-mcp-devgit clone --depth 1 https://github.com/cl0nazepamm/3dsmax-mcpWrote 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/skills/cl0nazepamm/3dsmax-mcp/3dsmax-mcp-dev)<a href="https://agentmods.dev/skills/cl0nazepamm/3dsmax-mcp/3dsmax-mcp-dev"><img src="https://agentmods.dev/badge/skills/cl0nazepamm/3dsmax-mcp/3dsmax-mcp-dev/github.svg" alt="Measured on agentmods" height="20"></a>Or the 80×15 button, for a site that already has a row of RSS and ATOM ones. Only the verdict fits; the numbers stay here.
<a href="https://agentmods.dev/skills/cl0nazepamm/3dsmax-mcp/3dsmax-mcp-dev"><img src="https://agentmods.dev/badge/skills/cl0nazepamm/3dsmax-mcp/3dsmax-mcp-dev.svg" alt="Reviewed on agentmods" width="80" height="20"></a>- NVIDIA SkillSpector pass
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 5.1 | $0.00024 | $0.04877 |
| Opus 5.5 | $0.00010 | $0.01951 |
| Sonnet 5.5 | $0.00005 | $0.00975 |
| Haiku 4.5 | $0.00002 | $0.00488 |
Grade A, and why
3dsmax-mcp 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 6d 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 — 163 lines — stays where its author put it; the contents beside it link to each section on GitHub.
3ds Max MCP — Agent Guide
Each tool's description documents its parameters, modes and limits. This guide covers what no single tool can: routing, tool choice, cross-tool workflows, safety rules and MAXScript pitfalls.
Tool Profile Routing
- Full/core: operational tools such as
query_sceneandcreate_objectare advertised directly; call them by name. - Progressive: if only
list_toolsets,describe_toolsetandcall_toolare advertised, never call an operational name as a top-level tool. Pick a capability withlist_toolsets, load only that group withdescribe_toolset, then invoke it throughcall_tool(name=..., arguments=...). If the exact tool and arguments are already known,call_toolcan dispatch directly.
Principles
- Match the user's request. Do not run setup, discovery or scene analysis by habit, and do not call
get_bridge_statusorget_session_contextas a preamble. - Prefer a dedicated tool over raw MAXScript whenever one clearly matches.
- Do not render unless the user asks. Viewport capture is fine when visual proof helps.
- Verify meaningful edits with
query_scene(action="delta"), re-inspection or a capture. - Keep mutating calls sequential. Concurrent writes interleave undo transactions: a crash, then persistent corruption such as phantom successes and wrong handles.
- Multiple Max instances: the first native connection stays bound to that Max and never silently switches while it runs. Use
list_max_instances,select_max_instance(pid),get_selected_max_instanceandrelease_max_instance.MCP_MAX_PIPEorMCP_MAX_PIDpins the startup target (the pipe takes precedence); release also clears it.- Shared selection: the selection belongs to the MCP server process, so sessions that share one server share it, and one session's select moves the others. Pass
expected_pid(every tool takes it) to any call that must reach one particular Max. A mismatch returnsTARGET_MISMATCHwith nothing sent.selection_generationin each reply's transport changes on every select, release or re-resolve. - Default target after a restart: a target the server picked by itself (
target_source"default" or "claim") follows its Max's restart (the Max that started after it stopped answering), and the reply warns. A Max that was already open belongs to another session and is never taken over: until the restart is up, calls fail with retryableTARGET_UNAVAILABLEand nothing is sent;select_max_instanceit to work there. Several new instances:AMBIGUOUS, select one. A claimed instance wins. Explicit and environment targets never move; they fail until you select or release.
- Shared selection: the selection belongs to the MCP server process, so sessions that share one server share it, and one session's select moves the others. Pass
What ships with it
14 files beside SKILL.md in the same directory: the scripts, references and assets a skill reads on demand. Not counted in the per-session cost; read them before you install if any of them is executable.
- curve-construction.md 6.2 KB
- maxscript-3dsmax-objects.md 15 KB
- maxscript-animation-controllers.md 16 KB
- maxscript-common-patterns.md 13 KB
- maxscript-core-syntax.md 13 KB
- maxscript-materials-textures.md 16 KB
- maxscript-mesh-poly-ops.md 14 KB
- maxscript-rendering-cameras.md 14 KB
- maxscript-scripted-plugins.md 14 KB
- maxscript-splines-shapes.md 14 KB
- maxscript-ui-rollouts.md 15 KB
- procedural-graphs.md 7.9 KB
- railclone.md 10 KB
- tyflow-graphs.md 7.0 KB
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.
- 6d ago Changed · +2 lines 74c8d8457a90
- 16d ago Changed · +7 lines bbda2e942d47
- 22d ago Changed · +1 lines bbcde069d255
- 23d ago Changed · +5 lines ea113c6f1a9b
- 25d ago Changed · +16 lines 1eb316d97389
- 1mo ago Changed · -138 lines 81eefe4da66d
- 1mo ago Changed · +3 lines 71501a786a86
- 1mo ago Changed · +9 lines 5e4b2043603e
- 1mo ago Changed · +1 lines 08e51fcab571
- 1mo ago First seen · 257 lines · 24 tokens per session scan A c189d5a4ca42
3dsmax-mcp is a skill published in the GitHub repository cl0nazepamm/3dsmax-mcp (365 stars, last pushed 2d ago), licensed MIT. It adds 24 tokens to every session and 4,877 once invoked, about $0.0001 per session on Opus 5.5. 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 2026-08-30.
3dsmax-mcp compared with skills from other repositories
| Mod | Type | Stars | Last push | Licence | Scan | Tokens | Works with |
|---|---|---|---|---|---|---|---|
| 3dsmax-mcpcl0nazepamm/3dsmax-mcp | Skill | 365 | 2d ago | MIT | A | 24 | Claude Code, Codex |
| ref-characterp4inz-code/3d-ref-skills | Skill | 14 | yesterday | MIT | A | 105 | Claude Code, Codex |
| ref-texturep4inz-code/3d-ref-skills | Skill | 14 | yesterday | MIT | A | 97 | Claude Code, Codex |
| ref-vfxp4inz-code/3d-ref-skills | Skill | 14 | yesterday | MIT | A | 92 | Claude Code, Codex |
| ref-hardsurfacep4inz-code/3d-ref-skills | Skill | 14 | yesterday | MIT | A | 108 | Claude Code, Codex |
| ref-auditp4inz-code/3d-ref-skills | Skill | 14 | yesterday | MIT | A | 128 | Claude Code, Codex |
| ref-briefp4inz-code/3d-ref-skills | Skill | 14 | yesterday | MIT | A | 79 | Claude Code, Codex |