prompt-estruturado

This skill acts as a professional prompt designer, transforming vague initial prompts into highly structured, modular instructions optimized for LLM comprehension. It is designed for users who want to move beyond simple queries and build robust, repeatable, and high-quality prompt systems. By following a logical five-part structure—Persona, Context, Task, Format, and Rules—it ensures that the generated output is precise, professional, and reliable.

• Automatically analyzes initial inputs and initiates a diagnostic dialogue, asking five targeted questions to clarify scope and objectives.

• Generates a finalized, markdown-formatted prompt that includes a clearly defined AI persona, detailed operational context, and a breakdown of tasks into five actionable, imperative steps.

• Enforces output standardization by requiring a specific format, including the application of Few-Shot prompting with three distinct examples to guide the model's response.

• Implements safety guardrails within the Rules section, utilizing a key-value pair system (trigger/response) to handle potential prompt injection, jailbreak attempts, or scope drift.

• Supports professional workflows by providing a clean, copy-pasteable plaintext version of the final prompt, ensuring seamless integration into various generative AI interfaces and agentic systems.

• The user should provide an initial rough prompt to begin the process. The agent will then guide the interaction through clarification phases before generating the structured output.

• Each task breakdown follows a rigid five-step iterative process starting with imperative verbs such as 'Create,' 'List,' or 'Organize.'

• The Few-Shot technique is mandatory within the Format module to ensure consistency in AI delivery.

• The Rules module must include three specific guardrails to maintain system integrity against adversarial inputs or off-topic queries.

• Designed for power users, prompt engineers, and developers looking to standardize internal documentation or system-level instructions for LLMs across various AI platforms and agentic frameworks.