Claude Code: Linux Operations Exercises

Practice exercises for Chapter 11: Linux Mastery for Digital FTEs

14 hands-on exercises across 3 tiers that take you from navigating server filesystems to diagnosing cascading production failures. Every exercise puts you in Ali's shoes — you direct Claude Code, read the output, and make decisions.

Package Structure

claude-code-linux-mastery-exercises/
├── EXERCISE-GUIDE.md                                # Full guide with rubrics and framework
├── README.md                                        # This file
├── tier-1-foundation/                               # Lessons 1-2: Navigation and reading
│   ├── exercise-1.1-server-orientation/             (Build: Explore a new server)
│   ├── exercise-1.2-mystery-server/                 (Debug: Find and diagnose an abandoned agent)
│   ├── exercise-1.3-reading-the-room/               (Build: Interpret server health output)
│   └── exercise-1.4-permission-puzzle/              (Debug: Diagnose permission errors)
├── tier-2-operations/                               # Lessons 3-5: Building infrastructure
│   ├── exercise-2.1-agent-home-setup/               (Build: Create project directory structure)
│   ├── exercise-2.2-missing-pieces/                 (Debug: Fix an intern's broken setup)
│   ├── exercise-2.3-service-from-scratch/           (Build: Write a systemd service)
│   ├── exercise-2.4-service-wont-start/             (Debug: Find 3 root causes in a failed service)
│   ├── exercise-2.5-lock-it-down/                   (Build: Harden a deployment's security)
│   └── exercise-2.6-overprivileged-agent/           (Debug: Security audit with 5 findings)
└── tier-3-diagnosis/                                # Lessons 6-7: Systematic debugging
    ├── exercise-3.1-silent-agent/                   (Debug: Agent running but not working)
    ├── exercise-3.2-cascading-failure/              (Debug: 3 agents down, 1 root cause)
    ├── exercise-3.3-deployment-spec/                (Build: Write a complete deployment spec)
    └── exercise-3.4-spec-vs-reality/               (Debug: Find gaps between spec and server)

Getting Started

Prerequisites

• Claude Code (required): All exercises involve directing Claude Code to investigate and fix Linux scenarios. Install following the instructions at https://docs.anthropic.com/en/docs/claude-code.
• No live server required: Exercises use simulated server output files. You practice reading and interpreting output, then directing Claude Code — no SSH or root access needed.

Setup

1. Download or clone this repository
2. Open a terminal in the repository root
3. Launch Claude Code: claude
4. Start with Tier 1 and work sequentially

Recommended Order

Work through tiers in order. Each tier builds on skills from the previous one.

Tier	Focus	Exercises	Time
Tier 1: Foundation	Server navigation, reading output, permissions	1.1 – 1.4	60-90 min
Tier 2: Operations	Directory setup, systemd, security hardening	2.1 – 2.6	90-120 min
Tier 3: Diagnosis	LNPS method, cascading failures, deployment specs	3.1 – 3.4	60-90 min

The LNPS Debugging Method

Tier 3 exercises use the LNPS triage method. Follow this order for every diagnosis:

1. Logs — What do the logs say? Read before guessing.
2. Network — Can the service reach what it needs? Check ports and connectivity.
3. Process — Is the process alive? What state is it in?
4. System — Are disk, memory, and CPU okay?

Exercise Workflow

For every exercise:

1. Read INSTRUCTIONS.md in the exercise folder
2. Examine the starter files — simulated server output, configs, logs
3. Direct Claude Code to investigate and diagnose
4. Create the deliverable (report, fix plan, audit, or spec)
5. Reflect using the questions in INSTRUCTIONS.md

Tips

• Read INSTRUCTIONS.md first in each exercise before opening Claude Code
• Never skip the investigation step — even when you think you know what is wrong
• For Debug exercises: read the broken state carefully before attempting any fix
• Reflect honestly — the reflection questions are where real learning happens