Author: Merkle3s Capital; Source: X,@Merkle3sCapital

Introduction

Code is dead, logic is immortal.

This is not alarmist. By 2026, AI will be able to easily translate JavaScript into Python, rewrite Go into Rust, and eliminate all traces of code plagiarism. But the isomorphism of architecture cannot be eliminated by rewriting—just like translating a book from Chinese to English, the story structure remains the same.

Chapter Two: What is the Hermes Agent?

How a top AI lab with $70M funding turned an "uncensored model" into a "7/24 self-evolving agent."

Nous Research: Who is building this agent?

Funding exceeds $70M (including a $50M Series A), 85,000 Stars on GitHub. Their most famous product is the Hermes LLM series—a large language model known for its "uncensored" nature; VentureBeat praised it as "outperforming ChatGPT without content restrictions." Now, what they're doing is transforming Hermes from a "model" into an "Agent." Core Positioning: 24/7 Autonomous Agent Unlike "programming assistants" like Claude Code and Cursor, Hermes Agent is positioned as a 24/7 autonomous agent running in the server background, similar to OpenClaw: Operation Mode: Server-resident, not IDE-bound. Evolution Capability: Learns through cycles, gets stronger with use. Memory: Persistent memory across sessions. left;">Platforms: Telegram/Discord/Slack/WhatsApp/Signal/WeChat/iMessage and 16 other messaging platforms

Three-tier memory architecture

The underlying support is an SQLite FTS5 extension, compared to OpenClaw's qmd Theoretically, this allows for faster memory retrieval. This design embodies the philosophy of "digital sovereignty"—all memory and execution logic remain in the user's local or private environment, complying with GDPR compliance narratives. 47 Built-in Tools with Anthropic Standards Web Search, Browser Automation, Terminal Execution, File Editing Image Analysis/Generation, Speech Synthesis, Code Execution Sandbox Sub-Agent Scheduling, Cron Tasks Hermes Agent adopts the **Agent Skills Open Standard** released by Anthropic in December 2025. This standard marks a shift in AI agents from monolithic intelligence to modular expertise. According to an industry survey at the end of 2025, the vulnerability rate of skill systems was as high as 26.1%—because it allowed the execution of arbitrary scripts. Hermes defends against sensitive information leaks through the agent/redact.py module. Chapter 3: What is Evolver? GEP License—an "AI Agent Evolution Middleware" open-sourced 36 days earlier than Hermes. Developed by the EvoMap team (AutoGame Limited), Evolver is positioned as "middleware that equips any AI Agent with evolutionary capabilities."

Core Concept: GEP Protocol

GEP (Genome Evolution Protocol) — A standardized evolutionary process, similar to gene expression in organisms:

• Gene: Reusable evolutionary assets, extracted from runtime logs

• Capsule: A higher-level evolutionary unit, encapsulating multiple related Genes

• Evolution Event: Each evolution has an audit trail

10-Step Evolutionary Cycle

1. Ensure Asset Files Exist

2. Three-Layer Signal Extraction (Regular Expression/Keyword/LLM)

3. Memory Map Consultation

4. Gene + Capsule Selection

5. Constructing Mutations

6. Selecting Personalities

7. Assembling GEP Prompts

8. Writing Prompt Files

9. Writing Solidified State

10. Reflection Assessment

4.1 Aligning the Four Core Modules One by One

Module 1: Generating a Closed Loop of Reusable Assets from Experience

Module 2: A Three-Layer Memory System

Module 3: Periodic Reflection Mechanism

Module 4: Skill Discovery and On-Demand Loading

4.2 Source Code Module One-to-One Correspondence

EvoMap Comment: Each Evolver core module is in Hermes There are functionally equivalent corresponding files in both.

4.3 Terminology Analysis: Systematic Concept Replacement

EvoMap performed a full-text search on the two Hermes repositories:

No direct code remnants were found.

5.1 Teknium's Denial

Teknium (co-founder of Nous Research, Head of Post-Training) responded on X:

I have never heard of them.

This is the core argument of this article, and also the true meaning of the title.

6.1 What is AI Codewashing?

7.3 "AI Codewashing" is a New Threat

This presents a new challenge to the open-source community: How to prove "AI-rewritten architecture plagiarism"?