Building With Llms

What it does

This skill, oldwinter-building-with-llms, enables AI agents to construct complex outputs by leveraging large language models (LLMs). It facilitates breaking down tasks into smaller steps and iteratively refining results. The agent can manage multiple LLM calls and combine their responses to achieve a desired outcome, offering greater control over the generation process.

When to use it

• Generating long-form content: When you need an AI agent to write articles, stories, or reports that require structured thinking and iterative refinement.
• Complex problem solving: For tasks involving multiple steps or dependencies where a single LLM call isn't sufficient.
• Creative writing with constraints: To guide the creative process by imposing specific rules or structures on generated text.
• Building interactive experiences: To create AI-powered applications that require dynamic content generation and user interaction.

Key capabilities

• Iterative LLM calls
• Task decomposition
• Response combination
• Structured output generation

Example prompts

• "Write a short story about a cat detective, breaking the process into outlining, drafting, and revising."
• "Create a marketing plan for a new product, first defining the target audience, then outlining key strategies, and finally detailing budget allocation."
• “Generate five different blog post titles related to sustainable living, then evaluate each one based on click-through potential.”

Tips & gotchas

The skill’s effectiveness depends on clear task instructions. Ambiguous prompts may lead to unpredictable or less desirable results; be specific about the desired output and process.