prompt-writing

This skill provides a robust framework for engineering professional-grade YAML prompts designed for AI agent orchestration. It is intended for developers and prompt engineers who need to maintain consistency, modularity, and scalability across their agent-based applications. By utilizing structured YAML patterns, this skill ensures that system prompts, user instructions, and dynamic variable placeholders are handled with precision, reducing hallucinations and improving model output reliability.

• Standardized YAML schema for system prompts and user interactions using Jinja2-style templating.
• Best practices for structuring multi-line prompts with clear section hierarchies using Markdown headers within YAML blocks.
• Quality assurance checklists to validate syntax, hierarchy, placeholder integrity, and separator usage.
• Reusable prompt design patterns including role definitions, specific requirements, guidelines, and constrained output formats.
• Integration support for agents built on orchestration platforms, ensuring compatibility with tools, MCP (Model Context Protocol), and memory management.

Usage involves defining roles, core responsibilities, and specific guidelines within YAML files to serve as consistent agent instructions. Users should leverage the provided checklist to verify that braces are closed, indentation is correct, and HTML tags are avoided to prevent parsing errors. When integrating with Jinja2, ensure that variables are clearly defined and descriptive to facilitate easy template instantiation. This skill is particularly effective when scaling projects from simple chatbots to complex autonomous agents that require reproducible, traceable, and highly structured input configurations.