Does this skill work across different programming languages?

Yes, it uses project-scoping to detect the environment (via Git) and applies the relevant instincts for the specific language or framework you are currently using.

How is my privacy protected during learning?

All observations and instincts are stored locally on your machine. The system only exports patterns (instincts), never raw code or full conversation history.

What is an 'instinct' in this system?

An instinct is a small, atomic learned behavior with a trigger, an action, and a confidence score that evolves based on session observations.

Why does it use hooks instead of standard Claude skills?

Standard skills are probabilistic and might not run every time. Hooks are deterministic, ensuring 100% of tool interactions are captured for a comprehensive learning record.

How do I promote a project-specific rule to all my projects?

You can use the /promote command or the /evolve command to move high-confidence patterns from a specific project directory to your global instinct library.

Continuous Learning & Instinct System v2

Name: Continuous Learning & Instinct System v2
Author: JantonioFC

byJantonioFC

•

Developer Tools

Learns and adapts to your unique coding patterns and project preferences through an atomic instinct-based architecture.

Continuous Learning v2.1 is an advanced knowledge-retention system for Claude Code that captures session activities and evolves them into reusable 'instincts.' By utilizing 100% reliable PreToolUse and PostToolUse hooks, the system identifies recurring patterns—such as preferred code styles, specific error-handling strategies, or repetitive workflows—and assigns them confidence scores. Version 2.1 introduces project-specific scoping to prevent pattern contamination between different repositories, allowing for hyper-specialized behavior in local projects while promoting universal best practices to a global knowledge base.

Key Features

01Global Promotion: Identifies high-confidence patterns appearing across multiple projects and suggests promoting them to a global scope.

02Evolution Pipeline: Automatically clusters related instincts to transform them into full-fledged skills, commands, or specialized agents.

03Project-Scoped Isolation: Keeps React patterns in React projects and Python rules in Python projects to prevent cross-contamination.

042 GitHub stars

05Atomic Instinct Generation: Captures specific behaviors with weighted confidence scores from 0.3 to 0.9.

06Deterministic Hook Integration: Uses PreToolUse/PostToolUse hooks for 100% reliable session observation compared to probabilistic methods.

Use Cases

01Standardizing Git workflows and input validation logic across multiple independent repositories.

02Maintaining consistent architectural patterns and component styles across a large-scale React or Vue application.

03Automating the capture of recurring debugging steps to create a personalized self-healing development environment.

What are Skills?·How to Install

Install with 🐟 Skill.Fish

npx skillfish add jantoniofc/skillsbank continuous-learning-v2

For use in Claude.ai and ChatGPT

Key Features

01Global Promotion: Identifies high-confidence patterns appearing across multiple projects and suggests promoting them to a global scope.

02Evolution Pipeline: Automatically clusters related instincts to transform them into full-fledged skills, commands, or specialized agents.

03Project-Scoped Isolation: Keeps React patterns in React projects and Python rules in Python projects to prevent cross-contamination.

042 GitHub stars

05Atomic Instinct Generation: Captures specific behaviors with weighted confidence scores from 0.3 to 0.9.

06Deterministic Hook Integration: Uses PreToolUse/PostToolUse hooks for 100% reliable session observation compared to probabilistic methods.

Use Cases

01Standardizing Git workflows and input validation logic across multiple independent repositories.

02Maintaining consistent architectural patterns and component styles across a large-scale React or Vue application.

03Automating the capture of recurring debugging steps to create a personalized self-healing development environment.