Enforces persistent user preferences and domain-specific constraints to prevent AI regressions across sessions.
The Frustration Eradication skill serves as a 'Layer 0' meta-enforcement layer for Claude Code, designed to solve the common issue of AI assistants failing to remember specific user preferences. By utilizing a ranked 'Frustration Index' and a Pre-Response Validation Checklist, the skill intercepts agent outputs to ensure compliance with mandates regarding tool selection (like Exa over web_search), response verbosity, coding styles, and specific voice parameters. It integrates with the Ruler protocol to propagate these constraints across multiple AI environments, ensuring a consistent, friction-free development experience that adheres to GF(3) triadic conservation principles.
Key Features
01Frustration Index for ranking and enforcing persistent user patterns
02Strict verbosity and preamble/postamble control
03Cross-agent configuration propagation via Ruler integration
048 GitHub stars
05GF(3) conservation logic for mathematical consistency in agent triads
06Pre-Response Validation Checklist to catch preference violations
Use Cases
01Ensuring AI responses remain concise and follow a strict maximum line count
02Maintaining strict code standards such as banning comments or 'as any' without repeated prompting
03Forcing the use of specific MCP tools like Exa for semantic search over default web tools
