The new Hermes Agent update officially ended OpenClaw

Introduction to Hermes Agent's New Features

Overview of the Update

• The latest update for Hermes Agent introduces several transformative features that enhance user experience and functionality.

• The video will cover seven new features, including how to effectively use Fable 5 with Hermes Agent.

Mixture of Agents: A Game-Changer

What is Mixture of Agents?

• Users can initiate a "mixture of agents" by typing /MA followed by a prompt, allowing multiple models to collaborate on responses.

• The aggregator model (Opus 48) sends prompts to reference models like ChatGPT and DeepSeek v4, which then return synthesized answers.

Benefits and Usage

• This feature results in higher quality responses as it combines insights from various models, tested against benchmarks showing improved outcomes.

• To utilize this feature, select "MOA default" from the dropdown menu and input your prompt for collaborative idea generation. An example prompt provided is about generating SaaS ideas based on personal interests.

Slasharn: Teaching Your Agent New Skills

Functionality of Slasharn

• The /slasharn command allows users to teach their agent new skills by providing URLs or prompts directly after the command. This automates skill creation based on external content like tweets or articles.

• For instance, users can take helpful tweets and convert them into skills simply by pasting the URL into Hermes with the /slasharn command.

Efficiency in Skill Development

• This method streamlines the process of enhancing agent capabilities without manual explanations or skill creation efforts, making it easier to self-improve your agent quickly.

Slashjourney: Visualizing Learning Progress

Overview of Slashjourney

• The /slashjourney feature provides a visual representation of all skills and memories accumulated by the agent over time through interactions with users. Users can see connections between different memories and skills developed during conversations.

Importance for User Engagement

• This tool helps track learning progress and ensures that agents are evolving alongside user needs, offering insights into what has been learned throughout usage sessions.

Cost-Efficient Self-Improvement

Enhancements in Cost Management

• With this update, every time an agent learns something via commands like /slasharn, it now utilizes cheaper models for implementing those learnings instead of relying solely on more expensive ones like Opus or ChatGPT models.

Implications for Regular Users

• Users should notice significant cost savings when using Hermes regularly due to these optimizations in resource allocation during self-improvement tasks within the platform.

Vibe Coding Improvements

Enhanced Vibe Coding Capabilities

• Major improvements have been made in vibe coding within Hermes desktop; users now have full Git controls integrated into their coding environment—allowing commits, pull requests, and code diffs directly from Hermes interface.

Practical Applications

• Demonstrations include building complex games such as action RPG-style projects entirely through vibe coding within Hermes Agent—showcasing its potential as a robust development tool compared to alternatives like Claude Code.

Utilizing Fable 5 with Hermes Agent

Setting Up Fable 5 Profiles

• Fable 5 is now available within Hermes; it's recommended that users create dedicated profiles specifically for utilizing this model effectively while managing costs associated with its use since it's not ideal as a daily driver due to expense concerns.(560]

Best Use Cases

• Fable 5 excels at handling complex tasks across devices or when detailed outputs are required (e.g., front-end development), whereas mixture of agents is better suited for synthesizing opinions on complex questions.(588]