Prompt Engineer

This is your reference guide when you need to write or fix prompts for Claude. It walks you through the full workflow: understanding requirements, picking techniques like chain of thought or multishot examples, structuring with XML tags, and adding quality controls for accuracy or security. The three reference files cover everything from basic clarity to advanced patterns like prompt chaining and long context handling. What's useful here is the troubleshooting matrix that maps common problems like hallucinations or inconsistent formatting directly to solutions. It's basically a checklist system to stop you from guessing which prompting technique to try next.

skills.sh installs: 171
GitHub stars: 234
last updated: Jun 3, 2026

View on GitHub

Install

$ npx -y skills add flora131/atomic --skill prompt-engineer --agent claude-code

Installs into .claude/skills of the current project.

Files

Prompt engineering

Create or revise prompts for the user's target model. Keep the common prompt portable and load only the relevant model guide. Each page connects an observed behavior to a prompt adjustment, prompt wording, and caveats. Use the matching patterns rather than pasting the whole guide into a prompt. The guides cover prompts, not API request settings; behavior does not transfer automatically between models.