book-prompt-optimizer

Automatically optimize AI remaster prompts by running them through a score-and-iterate loop with a linter and critic agent. This process systematically patches failing rules to ensure remaster outputs meet quality standards.

Batch processing of remastered sections is handled by running an automated optimization loop targeting linter rules. This iteratively scores and patches prompt instructions until the desired quality threshold is met.

Run the optimization process by executing the optimize_prompt.py script with arguments for your target score, iteration limit, provider, and model. For example, use the ollama provider flag with a specified model to start the loop.

A critic agent is required to patch failing linter rules during the prompt optimization process. The skill relies on both the linter and critic agent to execute the score-and-iterate loop for quality assurance.

You can set specific quality score targets and iteration limits by passing the target and iterations arguments to the optimization script. The loop runs until the target score is reached or the iteration limit is hit.

Systematic failures in AI remaster outputs often stem from poorly structured remaster prompts. This skill addresses these failures by automatically optimizing the prompt instructions using a linter as the scorer and a critic agent to patch rules.