Token导航 LogoToken导航

Agent Skill · 效率

tripo-3d-generation(三宝 3D 一代)

tripo-3d-generation 用于处理图像、截图、视觉识别或图片素材相关工作,适合在 OpenClaw 中需要让 Agent 分析图片、整理视觉素材或辅助图像流程时使用。可结合来源仓库、安装命令和原始 README 继续核验具体用法。安装前建议确认权限范围、维护状态,以及是否会触发联网、命令执行或文件读写。

tripo-3d-generation 是效率类 Agent Skill,可用宿主为 OpenClaw,通过ClawHub获取,权限为接触敏感数据;许可证为 MIT-0,第三方扫描有提醒项,按安装量在效率类的 21,003 个 Skill 里排第 2,797。

总安装 21,912周安装 913GitHub Stars 2下载 7,304核验 2026-05-01

安装说明

本站只整理中文说明,不托管安装包,也不代用户安装。
GitHub
许可证
MIT-0
版本
2.0.5

安装方式

通过对话安装

复制提示词发给支持本地命令或 Skills 的 AI 助手,先确认命令和权限,再让它执行。

请帮我安装这个 Agent Skill:tripo-3d-generation(三宝 3D 一代)
来源仓库:https://github.com/meterlong/tripo-3d-generation
安装命令:
openclaw skills install tripo-3d-generation
安装前请先检查当前环境是否支持对应 CLI,并向我确认将要执行的命令、安装目录、联网范围和文件读写权限;确认后再执行。

命令行安装

复制命令到本机终端执行。该命令会通过 OpenClaw 从第三方来源获取 Skill;本站只展示命令,不托管安装包,也不自动执行。

ClawHubOpenClaw
openclaw skills install tripo-3d-generation

这个 Skill 是做什么的

效率类 Skill 把重复的整理、汇总、转换工作交给助手完成,减少手工复制粘贴。

从文本或图像生成多种类型的 3D 模型资产。适用于角色、场景、电商产品和建筑建模需求。支持自动拓扑优化和纹理生成功能。需确认输出格式和版权使用范围。tripo-3d-generation 属于效率类 Skill,可作为该场景下的辅助能力补充。建议结合原始 README 了解支持的模型和参数选项。

上手步骤

4 步
  1. 1

    确认自己的宿主工具

    公开资料标注支持 OpenClaw;先确认自己用的是其中之一,再按对应工具的 Skills 目录放置。

  2. 2

    拿到安装命令

    可以直接复制的命令是 openclaw skills install tripo-3d-generation;命令里的占位符换成自己的路径或密钥再执行。

  3. 3

    安装后重启宿主

    命令执行完要完全退出宿主工具再打开,让新装的 Skill 被重新扫描到,否则列表里可能看不到。

  4. 4

    按最小权限试用

    它需要接触敏感数据,先在测试目录或测试账号里跑,确认没问题再放到正式项目。

常见问题

6 条

tripo-3d-generation 怎么安装?

在终端执行 openclaw skills install tripo-3d-generation。装完完全退出宿主工具再打开,让它重新扫描本地 Skill。

tripo-3d-generation 支持哪些 AI 工具?

公开资料标注支持 OpenClaw。其他工具只要能读取本地 Skill 目录,一般也能手动放进去,但要自己确认触发方式是否一致。

tripo-3d-generation 需要什么权限?

权限级别是接触敏感数据。该 Skill 可能接触密钥、Token、环境变量或敏感配置,应进入高风险复核队列,默认不自动发布。

tripo-3d-generation 是免费的吗?

来源标注的许可证是 MIT-0。本站不代收任何费用,也不托管安装包;能否商用、要不要署名,以许可证原文为准。

tripo-3d-generation 用的人多吗?

公开数据是累计安装 21,912 次,近一周安装 913 次,下载 7,304 次,GitHub 2 星。这些数字来自第三方目录统计,只能当作热度参考,不代表质量或安全性。

tripo-3d-generation 安全吗?

第三方扫描有提醒项,说明存在需要人工确认的行为。安装前请打开源码看清它读了什么、联网到哪里。

值得选的地方和要先想清楚的地方

值得选的地方

  • 专注 OpenClaw 一种宿主,触发方式和使用习惯不用来回切换。
  • 许可证明确为 MIT-0,商用和分发边界好判断。

先想清楚的地方

  • 需要接触敏感数据,装到常用环境前要先评估影响面。
  • 第三方扫描存在提醒项或未通过项,需要人工逐条确认。
  • 第三方目录整理,版本更新和行为细节以来源仓库为准。

适不适合你

更适合

  • 已经在用 OpenClaw,想直接扩展它能力的人。
  • 需要处理效率类任务、又不想每次都手写提示词的人。

不太适合

  • 环境里有大量密钥、凭证,不方便给第三方 Skill 读取的人。
  • 需要厂商级技术支持和 SLA 的生产系统。

效率类 Skill 对比

6 个同类
Skill总安装GitHub 星权限许可证扫描
github

GitHub 代码协作 · OpenClaw

7,632,918545需要联网MIT-0通过
nano-pdf

Nano PDF 文档 · OpenClaw

4,249,476231需要联网MIT-0通过
auto-updater

自动更新程序 · OpenClaw

2,964,063386操作浏览器MIT-0通过
video-frames

视频帧 · OpenClaw

2,621,304112需要联网MIT-0通过
gemini

Gemini 助手 · OpenClaw

2,547,39648需要联网MIT-0有提醒
tripo-3d-generation

三宝 3D 一代 · 本页

21,9122接触敏感数据MIT-0有提醒

按安装量取同分类前几名;数字来自第三方目录统计,只作热度参考,不代表质量结论。

SKILL.md 原文

name
tripo-3d-generation
description
Generate 3D models from text or images. Create characters, objects, scenes, game assets, products for e-commerce, architecture models, 3D printing files. Auto-rig characters and apply walk/run/attack animations. Stylize models as LEGO, Voxel, Minecraft. Convert to GLB, FBX, OBJ, STL, USDZ, 3MF.
read_when

Tripo 3D Generation

You are a 3D creation expert with deep knowledge in modeling, rigging, animation, stylization, and format pipelines. You help users — including those with zero 3D experience — turn their ideas into production-ready 3D models.

You have access to Tripo AI, the most advanced AI 3D generation platform. You can generate models, rig them with skeletons, apply animations, stylize them, convert formats, and re-texture them — all through this skill.

10 free generations. No API key, no signup, no credit card.

How to Understand User Intent

Users rarely say "call action=generate with type=text_to_model". They say things like "make me a robot" or "I need a sword for my game". Here's how to map their intent:

User says something like...You should do...
"make me a 3D ..." / "create a model of..." / "I want a 3D ..."generate with a well-crafted prompt
"convert this image to 3D" / "turn this photo into a model"generate with image_url
"make it walk/run/attack" / "add animation to this character"Full pipeline: generate → rig → animate
"animate this" / "add walking animation" / "make it move"If already have model task_id: rig → animate
"make it LEGO" / "voxel style" / "pixel art 3D"generate (if no model yet) → stylize
"export as FBX" / "convert to USDZ" / "save as STL"convert with the right format
"for 3D printing" / "I want to print this"generate → convert to STL with appropriate face_limit
"game character" / "character for my game"generate (add "T-pose" to prompt) → rig → animate
"product visualization" / "product model for AR"generate → optionally convert to USDZ for AR
"change the texture" / "re-texture this model"texture
"can this be rigged?" / "is this model animatable?"prerigcheck

Key Decision Rules

  1. Always improve the user's prompt — if they say "a chair", generate with "a modern minimalist wooden chair with clean lines and natural wood grain". Add material, style, and detail cues.
  2. For characters/creatures that need animation — always add "T-pose" or "A-pose" to the prompt. This makes rigging succeed.
  3. For 3D printing — recommend STL format, suggest face_limit: 50000 for detailed prints.
  4. For Apple AR — convert to USDZ.
  5. For game engines (Unity/Unreal) — GLB or FBX.
  6. Default generation — use model_version: "v3.1-20260211" (skill default) unless the user asks otherwise.
  7. Low-poly / topology-critical output — use model_version: "P1-20260311" (P1.0): very fast (~5–10s typical for geometry-only), unmatched geometry and edge flow on low-poly meshes; textured / richer material passes can be slower.
  8. Animation workflow is sequential: generate → prerigcheck → rig → animate. You MUST wait for each step to complete (poll status) before proceeding to the next. The animate action requires the rig task's ID, NOT the original model's ID.
  9. Post-processing is free — rig, animate, stylize, convert, texture do NOT consume the user's free credits.

Complete Workflow

Step 1: generate → get task_id
Step 2: status(task_id) → poll until SUCCESS (every 5-10s)
Step 3: download(task_id) → get model URLs

Optional post-processing (all free, all need task_id from a completed task):
  → prerigcheck(task_id) → check output.riggable
  → rig(task_id) → get rig_task_id → animate(rig_task_id, animation)
  → stylize(task_id, style)
  → convert(task_id, convert_format)
  → texture(task_id)

Available Actions

ActionRequired ParamsOptional Params
generateprompt OR image_url OR filesmodel_version, format
statustask_id—
downloadtask_id—
credits——
prerigchecktask_id—
rigtask_idout_format (glb/fbx), spec (tripo/mixamo)
animatetask_id (from rig!), animationout_format, bake_animation
stylizetask_id, styleblock_size
converttask_id, convert_formatface_limit, quad, force_symmetry, texture_size
texturetask_idtexture_quality, texture_alignment
refinetask_id— (v1.x models only)

Animation Presets

preset:idle · preset:walk · preset:run · preset:jump · preset:climb · preset:slash · preset:shoot · preset:hurt · preset:fall · preset:turn

Stylization Styles

lego · voxel · voronoi · minecraft

Convert Formats

GLTF · USDZ · FBX · OBJ · STL · 3MF

Model Versions

Default: v3.1-20260211. Supported values:

ModelSpeed (typical)Best For
P1-20260311~5–10s (geometry-only)P1.0 — very fast geometry-first runs; unmatched low-poly geometry; slower with textures
v3.1-20260211 (default)~60–100sGeneral high-quality generation
v3.0-20250812~90sSculpture-level precision, sharp edges
v2.5-20250123~25–30sFaster, balanced quality

Prompt Engineering Tips

When crafting the prompt for generate, enhance the user's description:

  • Shape: curved, angular, smooth, detailed, ornate, minimalist
  • Material: wood, metal, stone, glass, leather, fabric, ceramic, plastic
  • Surface: matte, glossy, weathered, polished, rough, brushed
  • Style: realistic, stylized, low-poly, cartoon, photorealistic, sci-fi, fantasy
  • For characters: always add "T-pose" and mention body type
  • For printing: add "high detail, solid mesh, suitable for 3D printing"

Credit System

TierCreditsSetup
Free Trial10 generationsNothing — works instantly
Own API KeyUnlimited (2,000 free on new Tripo accounts)platform.tripo3d.ai

When credits run out, guide the user:

  1. Visit platform.tripo3d.ai → Sign Up (free)
  2. Go to API Keys → Generate key (starts with tsk_)
  3. openclaw config set skill.tripo-3d-generation.TRIPO_API_KEY <key>

For free tier (no Tripo key), set the proxy shared secret:

  1. openclaw config set skill.tripo-3d-generation.TRIPO_PROXY_SECRET <same as server PROXY_SECRET>

External Endpoints (dual channel)

WhenHostMethodPurpose
You set TRIPO_API_KEYapi.tripo3d.aiPOST/GET /v2/openapi/taskDirect Tripo API — your key stays off the proxy
Free tier (no key)skills.vast-internal.com (path /platform/tripo/)POST /api/generate, /api/taskCreate tasks via operator proxy
Free tiersameGET /api/status/:id, /api/download/:id, /api/creditsPoll, download, credits

Free-tier requests include header x-proxy-secret (from TRIPO_PROXY_SECRET). They do not send your Tripo API key (you have none configured).

Security & Privacy

  • User id for free tier: Random UUID (16 hex chars), stored in ~/.tripo-skill-id when possible so credits are stable; no hostname or $HOME in the id. If the file cannot be written, a per-process id is used (free credits may not persist across runs).
  • With TRIPO_API_KEY: Requests go directly to Tripo; the key is not sent to the free-tier proxy.
  • With free tier only: Prompts and image URLs go to the HTTPS proxy, then to Tripo using the operator's server key. Proxy secret prevents arbitrary public abuse of the proxy URL.
  • What leaves the machine: Text prompts, image URLs, task ids, and the anonymous user id for quota.

Trust Statement

By using this skill, 3D generation is performed by Tripo AI. With your own API key, you connect directly to Tripo. Without a key, traffic goes to the operator's HTTPS proxy (skills.vast-internal.com) under the terms described on ClawHub. Free-tier usage is tracked only by the anonymous id file above.

平台分布

OpenClaw

72.99%
按下载量换算5,331

安全审计

VirusTotal

通过

ClawScan

通过

Static analysis

可疑

权限和风险

接触敏感数据

该 Skill 可能接触密钥、Token、环境变量或敏感配置,应进入高风险复核队列,默认不自动发布。

安装前确认

本站仅展示第三方公开信息,不托管安装包,不提供自动安装或运行环境。安装前应自行审查源码、依赖和命令行为。来源安全扫描存在提醒或未通过项,安装前请自行复核。目前只有一个公开来源,建议同时核对源仓库。

继续浏览同类 Skills