# Prompt Snippet Library A curated collection of reusable prompt snippets for system prompts, workflows, and AI agents. Each snippet includes both text and JSON integration examples. ## Repository Structure ### Model Identification - **model-identification.md** - Snippet for requesting model name and version identification ### Workflow Control - **single-turn-workflow.md** - Ensures definitive, complete responses without follow-up questions ### Formatting Instructions - **raw-markdown.md** - Raw markdown output formatting - **gmail-safe-html.md** - Email-client safe HTML formatting - **csv-output.md** - CSV data formatting within code blocks - **sql-output.md** - Valid SQL statement formatting ### Source Handling - **sources-section.md** - Dedicated sources section with simple hyperlinks ### Output Structure - **prompt-summary.md** - Begin output with improved summary of user's prompt - **verbatim-prompt.md** - Begin output with exact reproduction of user's prompt - **timestamp.md** - Add timestamp to start of every output ## Alphabetical Index - **CSV Output** - This snippet formats output as CSV data within a code block, producing immediately importable data that is compatible with Excel, Google Sheets, and database import tools. - **Gmail Safe HTML** - This Gmail-safe HTML formatting snippet is designed for email marketing templates and automated email generation, ensuring compatibility with major email clients while avoiding security issues and style stripping. - **Model Identification** - This prompt snippet is used to require users to identify their model version, facilitating logging, debugging, and route-specific workflows for models that support this functionality. - **Prompt Summary** - This prompt snippet helps generate a clear summary of the user's request, improving communication clarity and providing a structured output format for professional or documentation purposes. - **Raw Markdown** - This prompt snippet helps you generate clean and compatible raw markdown output for documentation systems like README files, reducing HTML injection risks and ensuring proper formatting. - **Single Turn Workflow** - This single-turn workflow ensures comprehensive, definitive responses without back-and-forth by providing clear instructions for model output, ideal for automated workflows with batch processing or API integrations. - **Sources Section** - This prompt snippet provides guidance on how to structure and integrate sources within written content, ensuring accessibility and readability across various systems. - **SQL Output** - This snippet generates immediately executable SQL code with proper syntax highlighting and protection against SQL injection, ideal for database administration and data analysis tasks. - **Timestamp** - This snippet provides a structured way to embed current date and time into user responses for logging, auditing, and time-sensitive purposes. - **Verbatim Prompt** - This prompt snippet ensures that user requests are accurately recorded and preserved, providing a clear audit trail for conversation logging, debugging, and maintaining accountability in automated systems. ## Usage Each snippet file contains: 1. **Text Version** - Ready-to-use prompt text 2. **JSON Integration Example** - How to integrate into JSON-based systems 3. **Usage Notes** - Context and best practices ## Integration Examples ### System Prompt Integration \`\`\`json \{ "system_prompt": "[snippet text] [your additional instructions...]", "user_prompt": "Your user prompt here" \} \`\`\` ### Workflow Chaining Combine multiple snippets for complex workflows: \`\`\` [Model Identification] + [Single Turn Workflow] + [Formatting Instructions] \`\`\` ## Contributing When adding new snippets: 1. Create appropriately named folders for categories 2. Include both text and JSON examples 3. Add usage notes and context 4. Update this README with the new snippet location