prompticorn 0.0.0

Build AI coding assistant configs from a shared prompt library

prompticorn

Version 0.1.0 - A unified, tool-agnostic prompt architecture for managing AI agent configurations across 5 coding assistants.

Define your agent once in a simple IR format, automatically generate outputs for:

• Kilo IDE (YAML + Markdown)
• Cline (Markdown with skill directives)
• Claude (Markdown .claude/ directory)
• Cursor (Markdown rules)
• GitHub Copilot (GitHub instructions)

Install

Install via pip:

pip install prompticorn

Or with uv:

uv add prompticorn

This installs the prompticorn CLI command.

Quick Start

See QUICKSTART.md for a 5-minute guide.

Key Features

• Unified IR System - Define agents once, generate for all tools
• 5 Production-Ready Builders - Kilo, Cline, Claude, Cursor, Copilot
• Minimal/Verbose Variants - Save tokens by choosing variant at build time
• Persona-Based Filtering - Select your team's roles, get only relevant agents
• Auto-Discovery Registry - Zero-config agent registration
• CLI Tool - Interactive prompticorn init command
• Backwards Compatible - Existing configurations still work

Commands Reference

Command	Description
prompticorn init	Interactively initialize prompt configuration for your project
prompticorn list	List all registered modes and their prompt files
prompticorn switch	Switch to a different AI assistant tool
prompticorn swap	Swap active personas and regenerate AI assistant configurations
prompticorn update	Update configuration options interactively
prompticorn validate	Check that all registered prompt files exist and no files are missing

Documentation

User Guides

• QUICKSTART.md - Quick start guide
• PERSONAS.md - Documentation on persona-based filtering
• GETTING_STARTED.md - Detailed getting started guide

Reference

• API_REFERENCE.reference.md - API reference documentation
• BUILDER_API_REFERENCE.builder.md - Builder API reference
• BUILDER_IMPLEMENTATION_GUIDE.builder.md - Guide for implementing new builders
• ARCHITECTURE_OVERVIEW.md - System architecture overview
• INDEX.md - Documentation navigation hub

Workflow

• Adding prompts: Run prompticorn init to generate configurations, then edit files in generated directories (e.g., .kilo/rules/)
• Updating prompts: Edit source prompts, then re-run prompticorn init to regenerate

Mode Reference

Mode	Key	Purpose
Architect	architect	Scaffold projects, task breakdowns, data models
Test	test	Coverage-first test writing
Refactor	refactor	Structural changes, behavior preserved
Document	document	Docstrings, READMEs, changelogs
Explain	explain	Code walkthroughs for onboarding
Migration	migration	Dependency upgrades, framework ports
Code	code	Feature implementation, boilerplate
Review	review	Code, performance, accessibility review
Debug	debug	Root cause, log analysis, rubber duck
Ask	ask	Q&A, decision logs
Security	security	Security review (code and infra)
Compliance	compliance	SOC 2, ISO 27001, GDPR, HIPAA, PCI-DSS
Orchestrator	orchestrator	CI/CD, DevOps, PR descriptions

Development

To contribute or develop locally:

# Clone the repository
git clone https://github.com/snoodleboot-io/prompticorn.git
cd prompticorn

# Install in development mode
pip install -e .

# Or with uv
uv pip install -e .

# Run tests
pytest -v

# Type checking
pyright