Skip to content

prompt-refinement

Domain: prompt · Model class: cheap

Use this skill when the user wants to work on Iteratively improving prompts based on measured failures and eval results. Triggers include “optimize this prompt”, “improve my prompt”, “my prompt keeps hallucinating”. Do NOT use when create a new prompt from scratch (use core-prompt-engineering).

Iteratively improving prompts based on measured failures and eval results. This skill provides structured guidance, references, and worked examples to help produce high-quality, actionable outputs.

  • “optimize this prompt”
  • “improve my prompt”
  • “my prompt keeps hallucinating”
  • “refine based on eval results”
  • “A/B compare prompt versions”
  • create a new prompt from scratch (use core-prompt-engineering)
  • run the initial evaluation (use core-prompt-evaluation)
  1. What is the user’s goal and current state?
  2. What constraints (time, team, compliance) apply?
  3. Are there existing artifacts (specs, code, benchmarks) to reference?
  • prompt asset
  • explicit output contract
  • failure handling
  • worked example or usage guidance

prompt-engineering · eval-prompt · eval-variance