instruction-improver

Analyze conversation history feedback patterns to generate CLAUDE.md improvements.

Updated Aug 23, 2026

Search Tags:#prompt engineering #ai #conversation analysis #instruction tuning #claudemd #feedback analysis

Ready to run

npx skills add https://github.com/acabreragnz/dotfiles --skill instruction-improver-acabreragnz

💬 Prompt for ChatGPT / Claude Chat▼

Please help me install this Agent Skill.
Skill: instruction-improver
Source: https://github.com/acabreragnz/dotfiles/tree/main/.claude/skills/instruction-improver
Command: npx skills add https://github.com/acabreragnz/dotfiles --skill instruction-improver-acabreragnz

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill includes scripts (resource) and references (resource) components.

What problem does it solve?

This Skill helps improve AI instructions by analyzing user feedback, identifying recurring mistakes and successes, and suggesting updates to AI configuration files (CLAUDE.md).

Core Features & Use Cases

Feedback Analysis: Scans conversation history for user frustration, corrections, praise, and successful outcomes.
Pattern Identification: Detects recurring issues and wins in user interactions.
CLAUDE.md Improvement: Generates specific, actionable suggestions to update CLAUDE.md files for better AI performance.
Use Case: After a series of interactions where the AI repeatedly misunderstands a specific type of request, this Skill can analyze those interactions and propose a new rule for CLAUDE.md to prevent future misunderstandings.

Quick Start

Analyze my conversation history for patterns and suggest improvements for my user-level CLAUDE.md file.

Frequently Asked Questions about instruction-improver

High-intent search queries and answers about installing and using this skill.

FAQPage Schema

How do I refine AI instructions from user feedback patterns?▼

To refine AI instructions from user feedback, this Skill scans conversation history to identify recurring mistakes and successes, then generates actionable improvement suggestions for your CLAUDE.md files.

How does conversation analysis identify recurring AI mistakes?▼

Conversation analysis identifies recurring AI mistakes by searching interaction logs for negative signals like user frustration and corrections, detecting repeated issues to propose new configuration rules.

What's the best way to improve CLAUDE.md files for better AI performance?▼

The best way to improve CLAUDE.md files is to analyze user interaction data for recurring issues and wins, translating those feedback patterns into specific, actionable configuration updates.

Do I need a specific tool to search conversation logs for instruction tuning?▼

Yes, you need the memex tool to search conversation logs for instruction tuning, as it is required to identify the negative and positive signals within your historical interaction data.

Can I use conversation feedback to prevent AI misunderstandings in future interactions?▼

Yes, you can use conversation feedback to prevent AI misunderstandings by analyzing past interactions for repeated errors and proposing new CLAUDE.md rules to correct that specific behavior.

What types of user feedback signals indicate a need for instruction tuning?▼

User feedback signals that indicate a need for instruction tuning include expressions of frustration, direct corrections, praise, and successful outcomes detected within the conversation history.