Can I manually edit the instincts it follows?

Yes, instincts are stored as simple Markdown files in the instincts folder, allowing you to manually refine, add, or delete learned behaviors.

How does Instinct Apply learn my patterns?

It works as part of the Homunculus framework, which observes your coding habits and stores them as structured markdown files in your local .claude directory.

What do the confidence levels in instincts mean?

Confidence levels (0.3 to 0.9+) determine how strictly Claude follows a pattern, ranging from tentative suggestions to mandatory rules.

Does this skill slow down Claude's response time?

No, the skill is optimized for efficiency. It uses a lightweight domain-check system to only read relevant instincts for the specific task at hand.

Instinct Apply

Name: Instinct Apply
Author: humanplane

byhumanplane

•

209

Productivity & Workflow

Automates the retrieval and application of learned coding patterns and personal preferences during active development.

About

Instinct Apply is a core component of the Homunculus framework that enables Claude to recall and act upon personalized development patterns. By scanning stored 'instincts'—structured markdown files containing triggers, domains, and confidence levels—this skill ensures that Claude adheres to your specific code styles, testing preferences, and git workflows without manual reminders. It intelligently surfaces domain-specific guidance at the start of tasks, allowing for a self-evolving development environment that adapts to your unique workflow and historical decisions.

Key Features

Automated pattern retrieval from personal and inherited instinct directories
Lightweight check system designed to minimize performance overhead
209 GitHub stars
Confidence-weighted decision making for applying learned behaviors
Context-aware filtering based on domains like code-style, testing, and git
Support for structured instinct files including metadata and evidence

Use Cases

Recalling preferred debugging strategies and testing frameworks during troubleshooting
Standardizing git commit messages and branch management based on historical behavior
Enforcing project-specific naming conventions and architectural patterns automatically

About

Key Features

Automated pattern retrieval from personal and inherited instinct directories
Lightweight check system designed to minimize performance overhead
209 GitHub stars
Confidence-weighted decision making for applying learned behaviors
Context-aware filtering based on domains like code-style, testing, and git
Support for structured instinct files including metadata and evidence

Use Cases

Recalling preferred debugging strategies and testing frameworks during troubleshooting
Standardizing git commit messages and branch management based on historical behavior
Enforcing project-specific naming conventions and architectural patterns automatically