Find your next open-source project

Search by name, purpose, or keyword — or paste a GitHub repository URL
All skills

prompt-engineer

Write, refactor, and evaluate LLM prompts, returning templates, output schemas, and test cases.

When to use it

You have a working prompt but inconsistent outputs or no way to measure accuracy, and need a structured rewrite with test cases.

For example

You say

Use prompt-engineer to rewrite my product-review sentiment prompt so it reliably returns Positive, Negative, or Neutral, and give me test cases to check it.

You get
  • Refactored sentiment prompt
  • Test cases and results
  • Usage settings

How it works

  • Iterates one change at a time and flags failure patterns when accuracy falls below 80% on the test set.
  • Covers few-shot, chain-of-thought, and structured JSON or function-calling output patterns.