Can I use this on existing prompt files?

Yes, you can provide a file path to any .md, .txt, or .yaml file, and the skill will perform an atomic replacement once the optimization is verified.

What happens if the prompt is already precise?

The initial verification phase will return a 'VERIFIED' status, and the skill will inform you that no changes are necessary, saving you from unnecessary iterations.

What is a verification loop in prompt engineering?

A verification loop is a process where a prompt is analyzed by a secondary AI agent to find flaws, optimized based on that feedback, and then re-analyzed until no further precision issues are detected.

How does this skill handle conflicting instructions?

It identifies contradictions (e.g., 'be brief' vs. 'be thorough') and either infers a priority based on the prompt's mission or asks the user to define a specific threshold.

Optimize Prompt Precision

Name: Optimize Prompt Precision
Author: doodledood

bydoodledood

•

开发者工具

Refines and disambiguates AI prompts through iterative verification loops to ensure absolute clarity and intent alignment.

Optimize Prompt Precision is a specialized prompt engineering skill for Claude Code designed to transform vague or contradictory instructions into airtight prompts. It employs a rigorous verification-first workflow that uses a secondary verifier agent to identify ambiguities, undefined terms, and edge case gaps. By running up to five iterative optimization cycles, the skill systematically resolves conflicts and adds necessary thresholds, ensuring that your prompts cannot be misinterpreted by the LLM. It intelligently balances automated inference with user clarification, making it an essential tool for developers building reliable agentic workflows.

主要功能

01Context-aware inference to define vague terms like 'concise' or 'standard'

02Atomic file replacement to ensure safe prompt updates

03Automated conflict resolution between contradictory instructions

048 GitHub stars

05Interactive clarification prompts for high-uncertainty decisions

06Multi-pass verification loops to detect hidden ambiguities

07Supports both local file paths and direct inline text optimization

使用场景

01Standardizing prompt terminology across a development team

02Hardening system prompts for production-ready AI agents

03Disambiguating complex multi-step coding instructions

What are Skills?·How to Install

Install with 🐟 Skill.Fish

npx skillfish add doodledood/claude-code-plugins optimize-prompt-precision

For use in Claude.ai and ChatGPT

Download Skill