Advanced
Infrastructure & Security
1 Week
Agentic Bash Security for Claude Code and Pi Agent
Source: IndyDevDan
Reduce production risk from AI coding agents with shell access using a five-level Bash security model.
Lesson media
Sign in to watch and study
What you will be able to do
- Evaluate and reduce the risk of agentic coding tools that can execute shell commands, from prompt guidance through no-Bash architectures.
What you will ship
Design a safer agent harness using layered controls: prompt policy, allowlists, sandboxing, approval gates, and no-Bash alternatives.
Who it is for
- AI engineers
- developers
- and technical leaders deploying agentic coding assistants.
Prerequisites
- Programming
- ML Basics
Enrollment
Pay per course
One-time enrollment for this class. Consumer price is taken from the live Stripe product when connected.
After payment is verified, the course appears on your dashboard. YouTube lessons and NotebookLM notebooks open after you sign in.