Does this require an external database?

While it provides the logic for session management, it is often used in conjunction with tools like DuckDB for persistent, time-traversable storage.

How does this improve AI-driven development?

It allows different AI tools to share a 'unified memory,' preventing context loss and errors when passing complex tasks between different agentic systems.

Which agents are supported by this skill?

The skill is designed to synchronize state between Goose (-1), Claude (0/Ergodic), and Amp (+1) agents.

What are B3 Posets and Bumpus sheaves used for?

These are mathematical frameworks used to model and order the relationships between different pieces of data across agent sessions, ensuring structural integrity.

What is Unified Reafference?

It is a specialized skill that creates a shared session database for multiple AI agents, allowing them to maintain a consistent state using topological data structures.

Unified Reafference

Name: Unified Reafference
Author: plurigrid

byplurigrid

•

데이터베이스 관리

Synchronizes multi-agent session data using B3 Poset structures and Bumpus sheaves to maintain shared state across AI agents.

Unified Reafference provides a sophisticated cross-agent session database layer that utilizes topological data structures, specifically B3 Posets and Bumpus sheaves, to harmonize state across different AI entities. By assigning distinct roles—Minus, Ergodic, and Plus—to agents like Goose, Claude, and Amp, it enables a cohesive memory and feedback loop (reafference) that allows multiple LLM instances to operate within a unified mathematical universe. This skill is essential for complex, long-running multi-agent workflows where data consistency, structural integrity, and collaborative memory are required to solve advanced engineering problems.

주요 기능

01Ergodic agent role assignment for balanced state transitions

028 GitHub stars

03Compatibility with DuckDB for time-traveling data persistence

04B3 Poset implementation for structured, partially ordered state management

05Advanced topological data modeling using Bumpus sheaves

06Cross-agent session synchronization for Goose, Claude, and Amp

사용 사례

01Implementing feedback loops (reafference) between different AI personas

02Managing complex state in multi-agent collaborative coding environments

03Ensuring data consistency across distributed agentic workflows

What are Skills?·How to Install

Install with 🐟 Skill.Fish

npx skillfish add plurigrid/asi unified-reafference

For use in Claude.ai and ChatGPT

Download Skill