Skip to main content

Command Palette

Search for a command to run...

Running Hermes Agent in YOLO Mode.

Or: Going to Bed while the AI Continues to Crunch.

Updated
5 min readView as Markdown
Running Hermes Agent in YOLO Mode.
B

Thank you for reading this post.

My name is Brian and I'm a developer from New Zealand. I've been interested in computers since the early 1990s. My first language was QBASIC. (Things have changed since the days of MS-DOS.)

I am the managing director of a one-man startup called Digital Core (NZ) Limited. I have accepted the "12 Startups in 12 Months" challenge so that DigitalCore will have income-generating products by April 2024.

This blog will follow the "12 Startups" project during its design, development, and deployment, cover the Agile principles and the DevOps philosophy that is used by the "12 Startups" project, and delve into the world of AI, machine learning, deep learning, prompt engineering, and large language models.

I hope you enjoyed this post and, if you did, I encourage you to explore some others I've written. And remember: The best technologies bring people together.

Abstract.

I set the YOLO environment variable so Hermes Agent automatically approves all my tool calls without prompting me for confirmation each time. This lets me:

  • Work faster,

  • Work uninterrupted, and

  • Sleep through the night without worries.

Attributions:

None for this post. ↗.


An Introduction.

YOLO Mode — "You Only Live Once" — is a setting I use with Hermes Agent that bypasses the usual interactive approval step. Normally, every time the agent wants to run a shell command, edit a file, or browse the web, it pauses and asks me for confirmation. With YOLO Mode enabled, those prompts are skipped entirely. The agent trusts me to know what I am doing, and I trust the agent to execute my instructions faithfully. It is a power-user feature designed for people who work iteratively and want fewer guardrails in the way.


The Big Picture.

I use Hermes Agent as my daily driver for development, system administration, and content creation. Every session involves many tool calls — writing files, installing packages, searching documentation, editing configurations. In the default mode, each of those calls triggers a confirmation prompt. Over the course of a typical session, those prompts add up to significant friction. YOLO Mode removes that friction entirely. I stay in my flow state, my commands execute immediately, and I get my work done faster. The trade-off is that I must be more deliberate about what I ask — there is no safety net asking "are you sure?" before a potentially destructive operation. For my workflow, that trade-off is well worth it.


Prerequisites.

  • Ubuntu 22.04 LTS or later,

  • Hermes Agent.


Updating my Base System.

  • From the terminal, I update my system:
sudo apt clean && \
sudo apt update && \
sudo apt dist-upgrade -y && \
sudo apt --fix-broken install && \
sudo apt autoclean && \
sudo apt autoremove -y

Always Applying YOLO Mode to Hermes Agent.

  • In the terminal, I open the .bashrc file with the Nano text editor:
sudo nano ~/.bashrc
  • I scroll to the bottom of the page (CTRL + END).

  • I paste (CTRL + V) the following, save (CTRL + S) the change, and exit (CTRL + X) the Nano text editor:

HERMES_YOLO_MODE=1
  • I refresh the terminal:
source ~/.bashrc

I run the Hermes Agent to ensure that it starts in YOLO Mode:

hermes

The Results.

After enabling YOLO Mode and launching Hermes Agent, I notice the difference immediately. Commands that used to pause for a confirmation now fly through without interruption. My terminal sessions feel faster and more fluid. I can chain together complex tasks — clone a repo, install dependencies, run a build, and push the results — without the agent stopping to ask permission at each step. The only confirmation I still see is the initial startup message acknowledging that YOLO Mode is active. From that point onward, every tool call is automatically approved. My workflow accelerates noticeably, especially during long coding or configuration sessions.

When combined with the Background Mode (/bg), YOLO evolves into a beast.


In Conclusion.

YOLO Mode is a simple environment variable that reshapes how I interact with Hermes Agent. It removes the confirmation bottleneck and lets me maintain a faster, more productive rhythm. I accept the increased responsibility that comes with fewer guardrails, and I find the speed gain more than compensates.

It has become a permanent part of my setup because I compose my prompts in VS Code where I use a document called Prompts_v0.5.x.txt. The patch number, shown as x, indicates how many versions of this file exist. These allow me to compose new prompts using elements from past constructs that worked. I do NOT vibe my way through a prompt. Instead, I carefully construct cues that are unambiguous to AI models, which improves my chances of success.

The following example is a real-world prompt. Paths have been shortened to protect the innocent. Also note the inclusion of /bg (Background Mode) at the very start of the prompt. This invokes a sub-agent to process my prompt. More importantly, "You can continue chatting — results will appear when done." This means I can use Background Mode again to run another sub-agent to process another prompt. During the day, I carefully compose multiple prompts that, in the evening (before I go to bed), I run in Background Mode. This causes multiple sub-agents to handle my perfectly produced prompts in YOLO Mode.

💡
/bg You will analyse the https://github.com/OpenBMB/MiniCPM repo and the https://huggingface.co/openbmb/MiniCPM5-1B-Base Hugging Face page. Pay special attention to any details relating to the pre-trained base model, how to download the model, how to fine-tune the model, and how to post-train the model. You will use the .Templates/AsciiDoc_Template_v0.5.1.ad template as your style guide as you create the very detailed ./Docs/Articles/MiniCPM5_Base_Model_v0.5.1.ad report that covers all the details relating to the MiniCPM5 AI model in general, and the MiniCPM5 base model in particular.

YOLO Mode works best with a proper prompting process.

Until next time: Be safe, be kind, be awesome, kia kaha!!


Hash Tags.

#YOLOMode #HermesAgent #Productivity #Automation #Workflow #PowerUser #Linux #Bashrc #EnvironmentVariable

The AI Series

Part 1 of 31

In the AI series, I describe numerous "artificial intelligence" technologies and techniques including machine learning, deep learning, prompt engineering, and large language models.

Up next

Installing Hermes Agent.

Or: Introducing the Best AI Agent for 2026.