allium

Generates executable specifications and integration tests from natural language descriptions of expected behavior, ensuring code aligns with stated requirements.

446|24|Updated Jan 25, 2026
One-click install
npx skills add https://github.com/juxt/allium --skill allium
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: allium
Source: https://github.com/juxt/allium/tree/main
Command: npx skills add https://github.com/juxt/allium --skill allium

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill includes scripts (resource) and references (resource) components.

What problem does it solve?

This Skill addresses the challenge of distinguishing between what code does and what it should do, providing a durable, discoverable form for behavioural intent that persists across conversations and sessions.

Core Features & Use Cases

  • Formal Specification: Define software behaviour using a precise, LLM-native language.
  • Intent vs. Implementation: Clearly capture what the system should do, separate from its current code.
  • Use Case: Generate integration tests, ensure code aligns with requirements, and facilitate clearer communication between developers and stakeholders by having a single source of truth for behavioural intent.

Quick Start

Use the allium skill to write a new rule for the 'User' entity that requires the user's email to be unique.

Frequently Asked Questions about allium

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I separate software behavioral intent from implementation details?

Software behavioral intent is separated from implementation by defining what a system should do using a formal specification language, creating durable models that persist across conversations independently of the current codebase.

What is the best way to distill behavioral specifications from existing code?

Behavioral specifications are distilled from existing code by extracting the underlying intent into an unambiguous, LLM-native formal language, ensuring the documented behavior aligns with developer velocity and clarity.

Can I generate integration tests from a formal behavioral specification?

Integration tests can be generated from a formal behavioral specification by using the precise intent models as a single source of truth, ensuring the test suite validates required system behavior rather than implementation quirks.

How do I capture software requirements through conversation elicitation?

Software requirements are captured through conversation elicitation by interacting with an LLM-native language to refine and distill ambiguous stakeholder requests into sound, maintainable behavioral models.

Does formal behavioral modeling work without modifying the underlying source code?

Formal behavioral modeling works without modifying source code by defining the expected behavior in a separate, durable formal language, allowing you to specify rules and intent independently of the implementation.