From 2a5415889920b4799865ecf06b321cf71091bd73 Mon Sep 17 00:00:00 2001
From: Alen Hu <huhaoyue0220@126.com>
Date: Thu, 16 Jul 2026 22:39:13 +0800
Subject: [PATCH] Add Modellix skill for AI image and video generation

Co-authored-by: Cursor <cursoragent@cursor.com>
---
 src/content/skills-zh/modellix.md | 41 +++++++++++++++++++
 src/content/skills/modellix.md    | 67 +++++++++++++++++++++++++++++++
 2 files changed, 108 insertions(+)
 create mode 100644 src/content/skills-zh/modellix.md
 create mode 100644 src/content/skills/modellix.md

diff --git a/src/content/skills-zh/modellix.md b/src/content/skills-zh/modellix.md
new file mode 100644
index 0000000..a369348
--- /dev/null
+++ b/src/content/skills-zh/modellix.md
@@ -0,0 +1,41 @@
+---
+slug: modellix
+---
+
+## 使用场景
+
+- 根据文本提示生成图片（T2I）
+- 编辑或改风格图片（I2I）
+- 从文本或图片生成视频（T2V / I2V）
+- 通过统一 CLI/API 调用 Modellix 模型，避免各厂商 SDK 割裂
+- 用「提交 → 等待 → 下载」一条龙流程自动化异步生成
+
+## 核心能力
+
+- **CLI 优先**：`doctor` → `model run --wait` → `task download`
+- **默认模型**：用户未指定 slug 时使用技能内默认值（如 T2I 用 Nano Banana lite，T2V 用 Seedance）
+- **REST 回退**：CLI 不可用时再走 HTTP API
+- **环境变量鉴权**：`MODELLIX_API_KEY`
+
+## 示例
+
+```bash
+# 预检
+modellix-cli doctor --json
+
+# 文生图（可省略 --model-slug 使用技能默认，或显式指定）
+modellix-cli model run \
+  --model-slug google/nano-banana-2-lite \
+  --body '{"prompt":"雨天街头的红色自行车，电影感"}' \
+  --wait --timeout 5m --json
+
+# 下载产物
+modellix-cli task download <task_id> --output-dir ./outputs --json
+```
+
+## 注意事项
+
+- 安装到 Qoder 时只复制 `modellix-skill/` 包目录到 `~/.qoder/skills/modellix`，不要复制整个仓库根目录。
+- 有 CLI 时优先用 `modellix-cli`，不要手写 REST 轮询。
+- 自动化场景优先使用 `--json` / `--quiet`。
+- 文档索引：https://docs.modellix.ai/llms.txt ；CLI：https://www.npmjs.com/package/modellix-cli
diff --git a/src/content/skills/modellix.md b/src/content/skills/modellix.md
new file mode 100644
index 0000000..434719f
--- /dev/null
+++ b/src/content/skills/modellix.md
@@ -0,0 +1,67 @@
+---
+name: modellix
+title: Modellix
+description: Generate images and videos via Modellix's unified model API and CLI (model run --wait, task download).
+source: community
+author: Modellix
+githubUrl: https://github.com/Modellix/modellix-skill/tree/main/modellix-skill
+docsUrl: https://docs.modellix.ai/llms.txt
+category: design
+tags:
+  - image-generation
+  - video-generation
+  - modellix
+  - cli
+  - api
+roles:
+  - designer
+  - developer
+  - marketer
+featured: false
+popular: false
+isOfficial: false
+installCommand: |
+  git clone https://github.com/Modellix/modellix-skill
+  cp -r modellix-skill/modellix-skill ~/.qoder/skills/modellix
+  npm i -g modellix-cli@latest
+date: 2026-07-16
+lastUpdated: 2026-07-16
+---
+
+## Use Cases
+
+- Generate images from text prompts (T2I)
+- Edit or restyle images (I2I)
+- Create videos from text or images (T2V / I2V)
+- Call Modellix models through one CLI/API instead of per-provider SDKs
+- Automate async generation with wait + download in a single workflow
+
+## Core Capabilities
+
+- **CLI-first workflow**: `doctor` → `model run --wait` → `task download`
+- **Default models** when the user omits a slug (e.g. Nano Banana lite for T2I, Seedance for T2V)
+- **REST fallback** when the CLI is unavailable
+- **Env auth**: `MODELLIX_API_KEY`
+
+## Example
+
+```bash
+# Preflight
+modellix-cli doctor --json
+
+# Text-to-image (omit --model-slug to use skill defaults, or set explicitly)
+modellix-cli model run \
+  --model-slug google/nano-banana-2-lite \
+  --body '{"prompt":"a red bicycle on a rainy street, cinematic"}' \
+  --wait --timeout 5m --json
+
+# Download artifacts
+modellix-cli task download <task_id> --output-dir ./outputs --json
+```
+
+## Notes
+
+- Install only the `modellix-skill/` package directory into `~/.qoder/skills/modellix` (not the whole monorepo root).
+- Prefer `modellix-cli` over hand-rolled REST polling when the CLI is available.
+- Prefer machine-readable flags (`--json` / `--quiet`) for automation.
+- See docs index: https://docs.modellix.ai/llms.txt and CLI: https://www.npmjs.com/package/modellix-cli