Visualizes and audits learned AI behavioral patterns to track confidence scores and domain distribution.
Instinct Status is a diagnostic tool designed for the Claude Code continuous learning ecosystem that provides a comprehensive overview of captured AI behaviors. It parses local instinct data to generate detailed reports on confidence scores, triggers, and evidence counts, grouping them by domain for better organization. By calculating key statistics and identifying clusters of related patterns, it helps developers recognize when a set of learned instincts has matured enough to be 'evolved' into a permanent, production-ready skill.
主な機能
01Comprehensive auditing of the local .claude/learning/instincts directory
02Clear visibility into instinct triggers, actions, and historical evidence
03Detailed statistical analysis of confidence levels and domain distribution
04Domain-based grouping for organized behavioral management
05Automated identification of evolution candidates for skill graduation
060 GitHub stars
ユースケース
01Auditing the 'instinct' database to ensure AI behaviors align with project standards
02Identifying high-confidence patterns that should be formalized into standalone skills
03Reviewing AI learning progress after a series of complex development sessions
