Favorite Agent Setups with Brian Christner

Bret Fisher Docker and DevOps
Bret Fisher Docker and DevOpsApr 14, 2026

Why It Matters

Secure, container‑based AI agents prevent operational failures and regulatory breaches, while skilled teams ensure businesses can harness AI’s productivity gains.

Key Takeaways

  • OpenClaw runs as root, causing catastrophic deletion risk.
  • Containerized AI agents like NanoClaw improve security isolation.
  • Companies adopt AI steering committees to enforce data protection policies.
  • Developers favor flexible models (OpenAI, Claude) over restrictive Copilot.
  • Rapid AI hype cycle pushes enterprises to treat AI as engineering tool.

Summary

The episode of the Agentic DevOps podcast features Bret interviewing Brian Christner, a Docker Captain and head of engineering for a major Swiss online casino. They explore current AI agent frameworks, focusing on OpenClaw, its successor NanoClaw, and the broader shift toward container‑based isolation.

Christner warns that running OpenClaw as root can lead to disastrous outcomes—he recounts a literal “delete everything” command that wiped his server. He argues that Docker‑native agents like NanoClaw, or VM‑based solutions such as Claude Cowork, provide the sandboxing needed to curb hallucinations and unintended system changes. The conversation also covers model selection, noting a split between easy‑to‑manage Microsoft Copilot and more flexible OpenAI or Claude APIs.

A standout anecdote describes the casino’s AI “enabler” who trains staff—from chefs creating AI‑generated brunch menus to developers building Kubernetes clusters—while a strict AI steering committee vets tools for data‑privacy compliance, rejecting Chinese models that violate EU regulations. The hosts also mention the cost volatility of fast‑mode LLMs, citing a $25 bill incurred in minutes.

Overall, the dialogue underscores that enterprises must adopt containerized AI agents, enforce governance frameworks, and upskill staff to avoid costly mishaps and maintain regulatory compliance as AI moves from autocomplete to a core engineering instrument.

Original Description

My friend Brian Christner (Docker Captain alum) and I go through our AI harnesses, agents, models, and what we’re playing with right now. OpenClaw, OpenCode, Claude Code, Copilot, and all of it.
Listen to the audio version of this show: https://agenticdevops.fm/episodes/our-favorite-agent-setups
This edited version is from my live stream show Mar 12, 2026: https://www.youtube.com/live/yh6zPMML4t8?si=wIn4_zyQZ9OiuJqp&t=229
Show Links
Brian’s Newsletter
Agents & Claude Code Setup
Claude Code Course
OpenClaw alternative
Brian's OpenClaw projects
🙌 I've launched the Agentic DevOps Guild, which is my premium community for accelerating your AI adoption for DevOps, CI/CD, platform engineering, and SRE. It includes courses, regular meetups, workshops, and mentorship. 🍾 https://www.bretfisher.com/theguild
🗞️ Sign up for my weekly newsletter for the latest on upcoming guests and what I'm releasing: https://www.bretfisher.com/newsletter/
Brian Christner
============
Bret Fisher
=========
Join my Community 🤜🤛
================
💌 Weekly newsletter on upcoming guests and stuff I'm working on: https://www.bretfisher.com/newsletter/
💬 Join the discussion on our Discord chat server https://discord.com/invite/devops
👨‍🏫 Coupons for my Docker and Kubernetes courses https://www.bretfisher.com/courses/
Chapters:
00:00 Introduction
03:01 AI Tools and Skills Deep Dive
30:52 Securing OpenClaw Servers
36:58 Safe Use Cases and Token Control
47:43 OpenClaw DOs and DON'Ts
53:11 NanoClaw and Container Future

Comments

Want to join the conversation?

Loading comments...