Hermes House

What Hermes House does

The README presents the project as an agent system that learns from each interaction, builds new skills and grows through self-improvement. Its architecture shows TrendRadar, Reviewer and Skills Manager agents sitting on an agent core with a memory system and a skills engine. Memory is plain files: MEMORY.md for core preferences and goals, tasks.md for tracking and a dated log per day.

Skills are grouped into development tools, data analysis, AI and ML, network and proxy, communication and business operations, with examples such as GitHub PR review, trend radar, SEO analysis, price tracking, Telegram and Twitter bots, content generation and backups. Monitoring panels cover TrendRadar, GitHub analytics and system health. You clone the repo, copy .env.example to .env and start with hermes start or python agent/main.py. The README reports 35 or more days of running and 51 or more skills, while the repository description claims 116 or more.

Key features

• TrendRadar monitoring agent and Reviewer agent
• Skill library grouped by category
• File-based memory with MEMORY.md, tasks.md and daily logs
• Monitoring for GitHub analytics and system health
• Examples for PR review, daily trend reports and content publishing
• Start through the Hermes CLI or a Python entry point

When to use it

• Running a daily trend report delivered to Telegram
• Having an agent review pull requests automatically
• Studying a worked example of a skill-driven Hermes setup

Who it is for: Chinese-reading builders who want an example of a large, self-extending Hermes-based agent setup.

How it fits with Hermes Agent

It is tagged hermes-agent and its quick start uses the hermes CLI, with the project presenting itself as an agent system built on that runtime.

How to install Hermes House

These commands are copied from the project's README. Check the repository for the latest steps before you run them.

git clone https://github.com/clowlove/hermes-house.git
cd hermes-house
cp .env.example .env
nano .env

Note: The repository description claims 116+ skills while the README counts 51+.

FAQ