What are hooks and why are they used for learning?

Hooks are deterministic triggers that fire every time a tool is used. This ensures the learning system captures 100% of relevant interactions, making it much more reliable than probabilistic skill-based observation.

Does this skill share my code or data externally?

No, all observations and learned instincts stay strictly local on your machine within your home directory, preserving your privacy and code security.

Can I share my learned behaviors with my team?

Yes, the skill includes /instinct-export and /instinct-import commands, allowing you to share specific patterns and behaviors without exposing raw conversation logs.

How does Continuous Learning differ from standard Claude memory?

Unlike basic memory, this skill uses atomic 'instincts' with confidence scoring and project-specific isolation, ensuring patterns only apply where they are relevant and avoid cross-project contamination.

Continuous Learning Instincts

Name: Continuous Learning Instincts
Author: Shawn-Guo-CN

byShawn-Guo-CN

0•

Productivity & Workflow

Automatically extracts and evolves reusable coding patterns and behaviors from your Claude Code sessions into permanent project-specific or global skills.

Continuous Learning v2.1 is an advanced instinct-based architecture that transforms transient Claude Code sessions into a persistent knowledge base. By utilizing deterministic hooks to observe every tool interaction, the system identifies repeated patterns—such as code style preferences, error resolutions, and specific workflows—assigning them confidence scores. It intelligently separates project-specific behaviors, like React patterns, from global best practices, like input validation, allowing your AI assistant to grow more personalized and efficient over time without cross-project contamination.

Key Features

01Deterministic observation using Pre/PostToolUse hooks for 100% reliable pattern capture

020 GitHub stars

03Intelligent promotion logic to elevate successful local patterns into global system-wide knowledge

04Confidence-weighted scoring system that evolves tentative observations into high-certainty behaviors

05Automated evolution of atomic instincts into complex reusable skills, commands, and agents

06Project-scoped instinct isolation to prevent conflicting coding styles between different repositories

Use Cases

01Onboarding to new projects by importing existing instinct libraries from other team members

02Reducing repetitive instructions by teaching Claude specific testing and git workflow preferences once

03Maintaining consistent architectural patterns across large-scale React or Python codebases

What are Skills?·How to Install

Install with 🐟 Skill.Fish

npx skillfish add shawn-guo-cn/agentfoundry-plugin-marketplace continuous-learning-v2

For use in Claude.ai and ChatGPT

Download Skill