# Create and Recover Quady3D Assets

Generating a 3D asset requires coordinating inputs, credit limits, task status, and file delivery. This skill guides Quady3D MCP workflows from authorized generation through recovery and project download.

## Install

```bash
npx skillstore add quady3dai/quady3d
```

## Metadata

- Status: approved
- Slug: quady3dai-quady3d
- Version: 0.1.2
- Author version: 0.1.2
- Skillstore revision: r1
- Version status: valid
- Tree hash: 6f9c40c317102d435d8d83ce64b54c6aa8c238f570ca745c9395d70d82b9df7d
- Author: quady3dai
- GitHub username: quady3dai
- License: MIT
- Repository: https://github.com/quady3dai/quady3d-tools/tree/13888045474eec1a0f6c2fff4a6205da7c12c4b0/skills/quady3d
- Ref: 3ce41f287227e078e7c84692ff9e43da7f23c094
- Supported tools: Claude, Codex, Claude Code
- Audit status: complete
- Agent install advisory: allowed
- Manual install advisory: allowed
- Artifact signature: available
- Audit attestation: unavailable
- Human verification: not\_verified
- Risk factors: env\_access, external\_commands, network
- Quality score: 85
- Quality tier: gold
- Public page: https://skillstore.pages.dev/skills/quady3dai-quady3d
- Manifest: https://skillstore.pages.dev/api/skills/quady3dai-quady3d/manifest

## Capabilities

- Query available models, parameter schemas, account limits, download access, and workspace readiness through Quady3D MCP.
- Prepare text input or upload one user-selected PNG, JPG, or WebP image up to 8 MiB.
- Request generation quotes and submit jobs using authorized credit limits and stable action identifiers.
- Recover interrupted submissions and query existing jobs without silently creating replacement generations.
- Download returned assets to authorized project paths while preserving existing files.
- Report saved paths, formats, sizes, and validation limits for delivered assets.

## Use Cases

- Prototype a Game Prop: Generate a single prop from text, save it into the game project, and check supported material, scale, and geometry requirements.
- Generate from a Selected Image: Upload an approved reference image, inspect model options, and generate a downloadable asset within an authorized credit limit.
- Recover an Interrupted Delivery: Find the original task, check its status, and retry asset delivery without starting a replacement generation.

## Prompt Templates

### Check Connection Readiness

```
Check Quady3D tools, account access, available text-generation models, and workspace readiness for [project directory]. Do not create a paid task.
```

### Generate a Text-Based Asset

```
Generate one [object description] for [intended use] in [project directory]. Keep spending within [maximum credits] and report the saved asset details.
```

### Use a Selected Reference Image

```
Use only [selected image path] to generate a 3D asset in [project directory] within [maximum credits]. Validate supported settings before quoting.
```

### Recover and Integrate an Existing Task

```
Recover [job or action ID] without creating a replacement. Download its successful asset into [project directory] and validate integration with the existing asset loader.
```

## Limitations

- Requires a separately installed Quady3D MCP server, Node.js 22 or later, credentials, credits, and appropriate download access.
- Only the documented standalone Codex workflow has been verified; Claude support remains unverified.
- Generation options depend on callable model capabilities; the skill does not independently provide rigging, animation, multiview generation, or editing.
- A downloaded model is not guaranteed to be printable, topologically sound, or compatible with project materials.

## Best Practices

- Keep API keys in secure host settings and restrict allowed roots to the intended project directories.
- Check model schemas, workspace readiness, download access, and authorized credit limits before creating a paid task.
- Preserve action, quote, job, and asset identifiers so interruptions can be recovered without duplicate generation.

## Anti Patterns

- Creating replacement generations after timeouts, stopped waits, or download failures without resolving the original task.
- Uploading unselected files, broadening project roots, or exposing credentials in conversation or project files.
- Claiming printability, rendering success, or advanced model features without relevant inspection and callable capabilities.

## Security Audit

- Audited at: 2026-11-05T20:45:09.057\+00:00
- Summary: All 16 static findings are false positives: Markdown identifiers, secure credential configuration, product links, and scoped workspace checks explain the matches. No evidence found of malicious intent, credential exfiltration, shell execution, or prompt injection in the reviewed files. The external MCP implementation and hosted service were not audited.

## Stats

- Views: 0
- Downloads: 1
- Favorites: 0
- Popularity score: 0