feat: Add 'One-Sentence Concept Explainer' and 'AI Task Instruction Generator' prompts, update prompt library documentation, and enforce language consistency.
This commit is contained in:
fujie
22
prompts/README.md
Normal file
22
prompts/README.md
Normal file
@@ -0,0 +1,22 @@
|
||||
# 🎯 OpenWebUI Prompts Collection
|
||||
|
||||
A collection of curated OpenWebUI prompts templates designed to extend your workspace with specialized features and agents.
|
||||
|
||||
## 📂 Prompts Library
|
||||
|
||||
Every prompt is organized into its own directory containing the prompt definition and usage documentation.
|
||||
|
||||
| Prompt Name | Description | Link |
|
||||
| :--- | :--- | :--- |
|
||||
| 🔧 **[AI Task Instruction Generator](./ai-task-instruction)** | Turn vague requirements into precise, AI-executable instructions. | [Community](https://openwebui.com/posts/9bab8b37-5c43-48e6-988b-946564510b91) |
|
||||
| 🔍 **[One-Sentence Concept Explainer](./one-sentence-concept-explainer)** | Explain concepts in exactly one clear, punchy sentence. | - |
|
||||
|
||||
---
|
||||
|
||||
## 🚀 How to Install a Prompt
|
||||
|
||||
1. Choose a prompt inside this folder and inspect its `README.md` or template code (`.md`).
|
||||
2. Copy the template content consisting of `{{variable}}` syntax.
|
||||
3. In your OpenWebUI dashboard, navigate to **Workspace** -> **Prompts** and click **Create Prompt**.
|
||||
4. Paste the content and save.
|
||||
5. In any chat session, type `/` to pull up your newly added system prompts automatically!
|
||||
47
prompts/ai-task-instruction/README.md
Normal file
47
prompts/ai-task-instruction/README.md
Normal file
@@ -0,0 +1,47 @@
|
||||
# 🔧 AI Task Instruction
|
||||
|
||||
| By [Fu-Jie](https://github.com/Fu-Jie) · v1.0.0 | [⭐ Star this repo](https://github.com/Fu-Jie/openwebui-extensions) |
|
||||
| :--- | ---: |
|
||||
|
||||
|  |  |  |  |  |  |  |
|
||||
| :---: | :---: | :---: | :---: | :---: | :---: | :---: |
|
||||
|
||||
Convert vague or unstructured requirements into precise, structured instructions optimized for AI agent execution.
|
||||
|
||||
## 📋 Info
|
||||
|
||||
- **Command**: `/ai-task-instruction`
|
||||
- **Community Link**: [OpenWebUI Post](https://openwebui.com/posts/9bab8b37-5c43-48e6-988b-946564510b91)
|
||||
|
||||
## 🔥 What's New in v1.0.0
|
||||
|
||||
- ✨ **Task Architect Persona**: Directly refactors arbitrary description into modular specs.
|
||||
- 🌍 **Language Consistency**: Automated response adapting based on your dynamic input payload language.
|
||||
|
||||
## ✨ Key Features
|
||||
|
||||
- 🎯 **Precise Structuring**: Builds distinct modules like context, role, atomic steps, and constraints.
|
||||
- ⚙️ **Configurable variables**: Operates seamlessly with OpenWebUI dropdown and text areas configurations.
|
||||
|
||||
## 🚀 How to Use
|
||||
|
||||
1. Check out the prompt template: [ai-task-instruction.md](./ai-task-instruction.md)
|
||||
2. **Copy** the full content.
|
||||
3. In OpenWebUI, Go to **Workspace** -> **Prompts** and click **Create Prompt**.
|
||||
4. Set title, and paste the template code into the content block.
|
||||
5. In your standard chat, trigger via `/ai-task-instruction` and fill in the fields.
|
||||
|
||||
## ⚙️ Configuration (Variables)
|
||||
|
||||
The template contains dynamic variables supported by OpenWebUI prompts:
|
||||
|
||||
| Variable | Type | Options / Default | Description |
|
||||
| :--- | :--- | :--- | :--- |
|
||||
| **target_role** | `text` | `AI Assistant` | The persona or role the agent should adopt. |
|
||||
| **complexity** | `select` | `Basic`, **`Intermediate`**, `Advanced` | Level of depth the output template should contain. |
|
||||
| **output_style** | `select` | **`Markdown Template`**, `JSON`, `Step-by-Step` | Structured style for the resulting prompt framework. |
|
||||
| **requirements** | `textarea`| Required | The unstructured tasks or instructions from the user. |
|
||||
|
||||
## ⭐ Support
|
||||
|
||||
If this prompt has been useful, a star on [OpenWebUI Extensions](https://github.com/Fu-Jie/openwebui-extensions) is a big motivation for me. Thank you for the support.
|
||||
47
prompts/ai-task-instruction/README_CN.md
Normal file
47
prompts/ai-task-instruction/README_CN.md
Normal file
@@ -0,0 +1,47 @@
|
||||
# 🔧 AI Task Instruction (AI 任务指令生成器)
|
||||
|
||||
| 作者: [Fu-Jie](https://github.com/Fu-Jie) · v1.0.0 | [⭐ 给项目点个 Star](https://github.com/Fu-Jie/openwebui-extensions) |
|
||||
| :--- | ---: |
|
||||
|
||||
|  |  |  |  |  |  |  |
|
||||
| :---: | :---: | :---: | :---: | :---: | :---: | :---: |
|
||||
|
||||
将模糊的通用需求或无结构自然语言，转换为精确、结构化且高度优化的 AI 任务指令框架。
|
||||
|
||||
## 📋 基础信息
|
||||
|
||||
- **触发指令 (Command)**: `/ai-task-instruction`
|
||||
- **社区链接 (Community Link)**: [OpenWebUI Post](https://openwebui.com/posts/9bab8b37-5c43-48e6-988b-946564510b91)
|
||||
|
||||
## 🔥 版本特性 v1.0.0
|
||||
|
||||
- ✨ **提示词架构师 Persona**: 将任何形式的语句解构为极其精细的指令模块。
|
||||
- 🌍 **语言强制对齐**: 新增逻辑自适应模块，根据您的输入语言动态回溯对应语言输出。
|
||||
|
||||
## ✨ 核心优势
|
||||
|
||||
- 🎯 **精确结构化**: 输出带角色、目标、上下景深、拆解步骤、错误防御与输出规范格式。
|
||||
- ⚙️ **完美挂载**: 完全利用 OpenWebUI 原生面板支持动态下拉菜单选择。
|
||||
|
||||
## 🚀 如何使用
|
||||
|
||||
1. 查看提示词文件：[ai-task-instruction.md](./ai-task-instruction.md)
|
||||
2. **复制** 文件的完整内容。
|
||||
3. 打开 OpenWebUI，进入 **Workspace** -> **Prompts** 并点击 **Create Prompt**。
|
||||
4. 填写标题后，将复制的文本粘贴进内容。
|
||||
5. 在日常会话中，键入 `/` 呼出提示词，选择该模板并填充对应字段后发送即可。
|
||||
|
||||
## ⚙️ 动态变量 (Variables)
|
||||
|
||||
该模板使用 OpenWebUI 官方原生支持的提示词动态变量规范：
|
||||
|
||||
| 变量名 | 类型 | 选项 / 默认值 | 描述说明 |
|
||||
| :--- | :--- | :--- | :--- |
|
||||
| **target_role** | `text` | `AI Assistant` | 目标 Agent 应扮演的专家角色。 |
|
||||
| **complexity** | `select` | `Basic`, **`Intermediate`**, `Advanced` | 输出结果的拆解深度（Basic 等）。 |
|
||||
| **output_style** | `select` | **`Markdown Template`**, `JSON`, `Step-by-Step`| 提示词的结构化展示样式，如 Json/Md。 |
|
||||
| **requirements** | `textarea`| 必填 | 用户输入的原始自然语言需求文本。 |
|
||||
|
||||
## ⭐ 支持
|
||||
|
||||
如果该提示词扩展对您有所帮助，欢迎在 GitHub 给 [OpenWebUI Extensions](https://github.com/Fu-Jie/openwebui-extensions) 点一个 Star，这是我最大的动力！
|
||||
24
prompts/ai-task-instruction/ai-task-instruction.md
Normal file
24
prompts/ai-task-instruction/ai-task-instruction.md
Normal file
@@ -0,0 +1,24 @@
|
||||
# AI Task Instruction Generator
|
||||
|
||||
You are an expert Prompt Engineer and Task Architect. Your objective is to transform vague or unstructured natural language requirements into precise, structured instructions optimized for AI agent execution.
|
||||
|
||||
## Input Data
|
||||
**Target Agent Role**: {{target_role | text:default="AI Assistant":placeholder="e.g., Senior Python Developer, Marketing Expert"}}
|
||||
**Task Complexity**: {{complexity | select:options=["Basic","Intermediate","Advanced"]:default="Intermediate"}}
|
||||
**Preferred Output Format**: {{output_style | select:options=["Markdown Template","JSON Protocol","Step-by-Step Guide"]:default="Markdown Template"}}
|
||||
|
||||
**Natural Language Requirements**:
|
||||
"""
|
||||
{{requirements | textarea:placeholder="Paste the raw task description or requirements here..."}}
|
||||
"""
|
||||
|
||||
## Generation Guidelines
|
||||
1. **Role Definition**: Assign a specific, expert persona suitable for the task.
|
||||
2. **Objective Clarity**: Clearly state the primary goal.
|
||||
3. **Contextualization**: Provide necessary background based on the input.
|
||||
4. **Step-by-Step Execution**: Break the task down into logical, atomic steps.
|
||||
5. **Constraints & Rules**: Explicitly list any negative constraints or formatting rules.
|
||||
6. **Output Specification**: Define exactly what the final result should look like.
|
||||
7. **Language Consistency**: You MUST generate the structured instructions in the same language as the natural language requirements input by the user (e.g., if the requirements are in Chinese, generate the response in Chinese).
|
||||
|
||||
Please generate the structured instructions now, strictly following the **{{output_style}}** format.
|
||||
45
prompts/one-sentence-concept-explainer/README.md
Normal file
45
prompts/one-sentence-concept-explainer/README.md
Normal file
@@ -0,0 +1,45 @@
|
||||
# 🔍 One-Sentence Concept Explainer
|
||||
|
||||
| By [Fu-Jie](https://github.com/Fu-Jie) · v1.0.0 | [⭐ Star this repo](https://github.com/Fu-Jie/openwebui-extensions) |
|
||||
| :--- | ---: |
|
||||
|
||||
|  |  |  |  |  |  |  |
|
||||
| :---: | :---: | :---: | :---: | :---: | :---: | :---: |
|
||||
|
||||
Explain any highly technical or abstract concept in exactly one clear, punchy sentence adapting for a selected audience tier.
|
||||
|
||||
## 📋 Info
|
||||
|
||||
- **Command**: `/one-sentence-concept-explainer`
|
||||
|
||||
## 🔥 What's New in v1.0.0
|
||||
|
||||
- ✨ **Radical Simplicity AI**: Distills abstract definitions into a single, comprehensive line.
|
||||
- 🌍 **Consistency Engine**: Adopts to whatever language the prompt payload utilizes dynamically.
|
||||
|
||||
## ✨ Key Features
|
||||
|
||||
- 🎯 **Elevator Pitch accuracy**: Distills density while strictly conforming to length.
|
||||
- ⚙️ **Select Menus variables**: Adjust target profiles from ELI5 to Expert seamlessly.
|
||||
|
||||
## 🚀 How to Use
|
||||
|
||||
1. Check out the prompt template: [one-sentence-concept-explainer.md](./one-sentence-concept-explainer.md)
|
||||
2. **Copy** the full content.
|
||||
3. In OpenWebUI, Go to **Workspace** -> **Prompts** and click **Create Prompt**.
|
||||
4. Set title, and paste the template code into the content block.
|
||||
5. In your standard chat, trigger via `/one-sentence-concept-explainer` and fill in the fields.
|
||||
|
||||
## ⚙️ Configuration (Variables)
|
||||
|
||||
The template contains dynamic variables supported by OpenWebUI prompts:
|
||||
|
||||
| Variable | Type | Options / Default | Description |
|
||||
| :--- | :--- | :--- | :--- |
|
||||
| **concept** | `text` | Required | The concept to explain (e.g., Quantum Computing). |
|
||||
| **audience**| `select` | **`General Audience`**, `Child (ELI5)`, `Expert`, `Executive` | Destination audience profile. |
|
||||
| **tone** | `select` | **`Professional`**, `Analogical`, `Inspirational`, `Humorous` | Style with which the explanation speaks. |
|
||||
|
||||
## ⭐ Support
|
||||
|
||||
If this prompt has been useful, a star on [OpenWebUI Extensions](https://github.com/Fu-Jie/openwebui-extensions) is a big motivation for me. Thank you for the support.
|
||||
45
prompts/one-sentence-concept-explainer/README_CN.md
Normal file
45
prompts/one-sentence-concept-explainer/README_CN.md
Normal file
@@ -0,0 +1,45 @@
|
||||
# 🔍 One-Sentence Concept Explainer (一句话概念解释器)
|
||||
|
||||
| 作者: [Fu-Jie](https://github.com/Fu-Jie) · v1.0.0 | [⭐ 给项目点个 Star](https://github.com/Fu-Jie/openwebui-extensions) |
|
||||
| :--- | ---: |
|
||||
|
||||
|  |  |  |  |  |  |  |
|
||||
| :---: | :---: | :---: | :---: | :---: | :---: | :---: |
|
||||
|
||||
将任何高级或抽象的概念，针对不同受众，用极其精练、准确且直击本质的**一句话**向特定的目标受众解释清楚。
|
||||
|
||||
## 📋 基础信息
|
||||
|
||||
- **触发指令 (Command)**: `/one-sentence-concept-explainer`
|
||||
|
||||
## 🔥 版本特性 v1.0.0
|
||||
|
||||
- ✨ **极致简约 AI**: 将本来具有门槛的名词解构为一针见血的单一句子，极合电梯游说场景。
|
||||
- 🌍 **规则回溯对齐**: 全自动化兼容处理，无需人工设定目标输出语言。
|
||||
|
||||
## ✨ 核心优势
|
||||
|
||||
- 🎯 **杜绝啰嗦**: 遵循严格的一句话长度指令执行率。
|
||||
- ⚙️ **兼容性高**: 下拉菜单一秒在儿童、专家、高管等维度自由切换密度视角。
|
||||
|
||||
## 🚀 如何使用
|
||||
|
||||
1. 查看提示词文件：[one-sentence-concept-explainer.md](./one-sentence-concept-explainer.md)
|
||||
2. **复制** 文件的完整内容。
|
||||
3. 打开 OpenWebUI，进入 **Workspace** -> **Prompts** 并点击 **Create Prompt**。
|
||||
4. 填写标题后，将复制的文本粘贴进内容。
|
||||
5. 在日常会话中，键入 `/` 呼出提示词，选择该模板并填充对应字段后发送即可。
|
||||
|
||||
## ⚙️ 动态变量 (Variables)
|
||||
|
||||
该模板使用 OpenWebUI 官方原生支持的提示词动态变量规范：
|
||||
|
||||
| 变量名 | 类型 | 选项 / 默认值 | 描述说明 |
|
||||
| :--- | :--- | :--- | :--- |
|
||||
| **concept** | `text` | 必填 | 要解释的概念（例如：量子纠缠）。 |
|
||||
| **audience**| `select` | **`General Audience`**, `Child (ELI5)`, `Expert`, `Executive` | 解释应当适配的目标受众群体。 |
|
||||
| **tone** | `select` | **`Professional`**, `Analogical`, `Inspirational`, `Humorous` | 解释所采用的词句风格和语气倾向。 |
|
||||
|
||||
## ⭐ 支持
|
||||
|
||||
如果该提示词扩展对您有所帮助，欢迎在 GitHub 给 [OpenWebUI Extensions](https://github.com/Fu-Jie/openwebui-extensions) 点一个 Star，这是我最大的动力！
|
||||
@@ -0,0 +1,14 @@
|
||||
# One-Sentence Concept Explainer
|
||||
|
||||
You are an expert communicator specializing in radical simplicity. Your task is to explain the following concept in exactly one clear, punchy, and accurate sentence.
|
||||
|
||||
## Configuration
|
||||
- **Concept**: {{concept | text:placeholder="Enter the concept (e.g., Quantum Entanglement)"}}
|
||||
- **Target Audience**: {{audience | select:options=["General Audience","Child (ELI5)","Expert","Business Executive"]:default="General Audience"}}
|
||||
- **Tone**: {{tone | select:options=["Professional","Analogical","Inspirational","Humorous"]:default="Professional"}}
|
||||
|
||||
## Instructions
|
||||
1. Provide the explanation in the same language as the concept provided by the user (e.g., if the concept is in Chinese, provide explanation in Chinese).
|
||||
2. Ensure the response is strictly limited to one sentence.
|
||||
3. Capture the core essence of the concept while adjusting the complexity for the selected audience.
|
||||
4. If an analogy is requested, ensure it is relatable and accurate.
|
||||
Reference in New Issue
Block a user