How does this skill utilize MCP Tasks?

It maps the agent's modification process to the Model Context Protocol (MCP) task specification, using states like 'working' for active changes and 'input_required' for human-in-the-loop verification.

What role does the Narya bridge play in this skill?

Narya provides structure-aware version control, treating code diffs as logical relations and managing complex conflicts through cubical geometry to ensure modification integrity.

How is cognitive continuity maintained during rewriting?

The skill uses Babashka transients and state-forking mechanisms to track the agent's 'cognitive state,' ensuring that the intent and context are preserved even as the underlying code changes.

What is Lisp-machine-style self-modification?

It refers to a computing paradigm where the software can modify its own source code and behavior during execution, allowing for highly dynamic and adaptive AI agents.

Codex Self-Rewriting

Name: Codex Self-Rewriting
Author: plurigrid

byplurigrid

•

開発者ツール

Enables Lisp-machine-style self-modification for AI agents using MCP Tasks and structure-aware Narya bridge types.

概要

The Codex Self-Rewriting skill empowers AI agents like Claude and Codex with advanced recursive modification capabilities, emulating the dynamic flexibility of classic Lisp machines. By leveraging the Model Context Protocol (MCP) and Narya observational bridges, it provides a formal framework for agents to safely modify their own logic and maintain cognitive continuity through Babashka transients. This skill is particularly useful for developers building self-evolving AI systems that require rigorous structure-aware version control and cross-editor state consistency through bisimulation equivalence.

主な機能

Inductive logical relations for complex diff and conflict resolution
MCP Task-based lifecycle management for agent self-modification
2 GitHub stars
Cognitive continuity via Babashka transient state tracking
Structure-aware version control using Narya observational bridge types
Cross-editor skill synchronization via bisimulation equivalence

ユースケース

Managing complex topological code modifications using formal structural relations
Maintaining consistent agent behavior across Claude, Codex, Cursor, and Copilot
Implementing autonomous agent self-evolution and logic patching

概要

主な機能

Inductive logical relations for complex diff and conflict resolution
MCP Task-based lifecycle management for agent self-modification
2 GitHub stars
Cognitive continuity via Babashka transient state tracking
Structure-aware version control using Narya observational bridge types
Cross-editor skill synchronization via bisimulation equivalence

ユースケース

Managing complex topological code modifications using formal structural relations
Maintaining consistent agent behavior across Claude, Codex, Cursor, and Copilot
Implementing autonomous agent self-evolution and logic patching