taste-learning

To make AI learn your preferences, this Skill observes your repeated corrections and edits to identify patterns in tone, length, and formatting, then injects these patterns as instructions into future runs.

Preference learning for adaptive AI systems is the process of detecting patterns in user output modifications to refine behavior automatically, ensuring future responses align with specific formatting and content inclusion rules.

You automate instruction injection by using this Skill to translate your repeated output modifications into actionable instructions, which are then automatically applied to refine future AI-generated content.

Yes, you can use preference learning to adjust output length and tone automatically, as the Skill specifically detects patterns in these modifications to refine AI behavior for future responses.

No, taste-learning does not require manual configuration for every preference, as it automatically detects patterns from repeated corrections and translates them into instructions without manual setup.

A limitation of detecting patterns from user feedback is that the Skill requires repeated corrections to identify preferences, meaning one-off edits may not be learned or applied to future AI outputs.