YOLO Developer Workflows with a Coding Agent in a Box
summary
Docker builds a lightweight microVM sandboxes (sbx) that run coding agents root‑inside but isolate host disk, network egress and secrets; demo shows sbx can publish ports, block API endpoints, and replace secrets via a secret manager so agents see fake tokens. Kits provide kits as declarative overlays (YAML) that persist tools, env and network policies; agents can run containers inside (docker‑in‑docker) to keep tests deterministic. Product is free but not yet open source and is controlled via CLI like sbx run / sbx secret.
AI-written summary · about
chapters (11)
- 0:00Introduction to YOLO workflows
- 1:32Autonomy vs safety in agents
- 6:13Need for sandboxing agents
- 8:23Introducing Docker Sandboxes
- 10:48Live demo of sandboxing
- 16:43Kits for sandbox customization
- 18:52Demo of kit installation
- 23:01Sandbox limitations and AI governance
- 24:45Speed vs safety trade-offs
- 27:18Installation and setup
- 29:49Q&A session
related talks
52mYou Gotta Keep the Dogs Away: YOLO Developer Workflows with a Coding Agent in a Box
33mYour Coding Agents in YOLO Mode… Safely
12mDocker Sandboxes for AI
26mrm -rf: Horror Stories From Unsandboxed AI Agents (and How Docker Fixes This)
27mFantastic Vibes and Where To Find Them
25mPlease Sandbox your AI Agents
01h 21mShip Agents that Ship: A Hands-On Workshop
34mZero Secrets: Building Secure Infrastructure for AI Agents