What is Unified Reafference?

It is a Claude Code skill that creates a shared session database, allowing different AI agents like Claude, Goose, and Amp to maintain consistent state and context during a project.

Why is 'reafference' important in AI workflows?

Reafference refers to the feedback loop of an agent's own actions. By unifying this across agents, the system ensures that every agent in the loop is aware of the state changes caused by the others.

Which database technologies does this skill support?

While it focuses on the topological structure of data, it is designed to work alongside tools like DuckDB for time-traveling data persistence and session narratives.

How does the B3 Poset structure work?

The B3 Poset organizes agents into specific roles: Goose acts as the 'Minus' (-1), Claude as the 'Ergodic' (0), and Amp as the 'Plus' (+1), creating a balanced mathematical framework for data exchange.

Unified Reafference

Name: Unified Reafference
Author: plurigrid

byplurigrid

•

Database Management

Synchronizes session data across multiple AI agents using a topological B3 Poset database structure.

About

Unified Reafference is a sophisticated cross-agent session management skill that enables seamless data persistence and state synchronization between Claude, Goose, and Amp. By leveraging mathematical frameworks like B3 Posets and Bumpus sheaves, it creates a 'topological chemputer' environment where agents can share context and feedback loops (reafference). This skill is essential for developers building complex, multi-agent orchestrations where maintaining a unified memory and state across different LLM instances is critical for consistency and logical flow.

Key Features

Bumpus sheaves integration for topological data consistency
Native support for DuckDB time-travel and narrative auditing
Ergodic session handling for continuous agent coordination
2 GitHub stars
Cross-agent session database synchronization for Goose, Claude, and Amp
B3 Poset-based state management with Trit-based logic roles

Use Cases

Maintaining a unified session history for auditing complex AI-to-AI interactions
Managing state in multi-agent 'topological chemputer' research and experiments
Synchronizing context between disparate AI agents during long-running development tasks

About

Key Features

Bumpus sheaves integration for topological data consistency
Native support for DuckDB time-travel and narrative auditing
Ergodic session handling for continuous agent coordination
2 GitHub stars
Cross-agent session database synchronization for Goose, Claude, and Amp
B3 Poset-based state management with Trit-based logic roles

Use Cases

Maintaining a unified session history for auditing complex AI-to-AI interactions
Managing state in multi-agent 'topological chemputer' research and experiments
Synchronizing context between disparate AI agents during long-running development tasks