What is a 'Trit' in the context of this skill?

The skill is part of a GF(3) triad where it acts as 'Trit 0'—the central coordinator that handles inference between expression validation and result generation.

How does color-based inference work in this context?

Using the !abduce command, the skill can take RGB values and search an 'invader space' to identify the specific ID or world state that corresponds to those colors based on topological models.

Which programming languages does this skill integrate with?

The skill integrates with Julia (specifically the Gay.jl package), Hy (using HyJAX), Babashka (clojure scripting), and Ruby for world state properties.

What is the primary purpose of the Abductive REPL skill?

It enables exploratory abductive reasoning, allowing the user to provide observations (like color data) and have Claude generate, test, and refine hypotheses to find the most likely cause or state.

Can I run automated tests with this skill?

Yes, it includes commands like !test and Justfile recipes for running abductive roundtrip tests to verify that inferred causes correctly reproduce observed effects.

Abductive REPL

Name: Abductive REPL
Author: plurigrid

byplurigrid

•

データサイエンスとML

Facilitates exploratory abductive inference through iterative hypothesis-test loops within a multi-language REPL environment.

Abductive REPL is a specialized Claude Code skill designed for exploratory reasoning, allowing users to infer underlying causes from observed effects via an interactive loop. By integrating with Julia (Gay.jl), HyJAX, and Babashka, it automates the process of generating hypotheses—such as mapping RGB color values to invader world states—and validating them through rigorous testing. This skill is ideal for researchers and developers working with topological models, complex system debugging, or any domain where observation-driven inference and iterative refinement are required to navigate high-dimensional state spaces.

主な機能

01Interactive hypothesis-test loops for exploratory abductive reasoning

02Multi-language integration supporting Julia, Hy, Babashka, and Ruby

03Automated abductive roundtrip testing and state verification

047 GitHub stars

05Specialized RGB-to-ID inference for topological invader spaces

06GF(3) triad coordination for validated expression and evaluation

ユースケース

01Reverse-engineering system logic from observed output patterns

02Exploring topological 'invader space' via color-based inference

03Automated validation of inference chains in complex state machines

What are Skills?·How to Install

Install with 🐟 Skill.Fish

npx skillfish add plurigrid/asi abductive-repl

For use in Claude.ai and ChatGPT

Download Skill