Troubleshoot Linux Systems Systematically

A 7-phase Linux troubleshooting workflow that pairs each diagnostic phase with named sub-skills and concrete commands. Phase 1 (Initial Assessment) invokes bash-linux and devops-troubleshooter to check uptime, review recent changes, and gather error messages via uptime, hostnamectl, cat /etc/os-release, and dmesg | tail -50. Phase 2 (Resource Analysis) invokes performance-engineer to check CPU, memory, disk, and I/O via top, free -h, df -h, and iostat -x 1 5. Phase 3 (Process Investigation) invokes server-management to find resource-hogging processes and inspect them via ps aux --sort=-%cpu, pstree -p, lsof -p PID, and strace -p PID. Phase 4 (Log Analysis) invokes error-detective to search system and application logs via journalctl -xe, tail -f /var/log/syslog, and grep -i error /var/log/*. Phase 5 (Network Diagnostics) invokes network-engineer to check interfaces, connections, and DNS via ip addr show, ss -tulpn, curl -v, and dig. Phase 6 (Service Troubleshooting) invokes systematic-debugging to check and restart services via systemctl status, journalctl -u <service> -f, and systemctl restart. Phase 7 (Resolution) invokes incident-responder and bash-pro to implement the fix, verify it, monitor stability, and document the solution and a prevention plan. Each phase ships a copy-paste @skill-name prompt to trigger the relevant sub-skill.

Use when diagnosing Linux system performance issues, service failures, network problems, disk space exhaustion, or application errors on a live system. Not for OS-level scripting or bash automation work with no active problem to diagnose - that's the sibling os-scripting and bash-scripting bundles - or for broader DevOps and cloud infrastructure work beyond a single host, which routes to cloud-devops.

Input is the observed symptom on a Linux system: slow performance, a failed service, a network issue, low disk space, or an application error. Output is a completed troubleshooting checklist (system info gathered, resources analyzed, logs reviewed, network tested, services verified, issue resolved, documentation created) plus quality gates confirming root cause identification, a verified fix, monitoring in place, and complete documentation.

top -bn1 | head -20
free -h
df -h
iostat -x 1 5

Orchestrates named sub-skills bash-linux, devops-troubleshooter, performance-engineer, server-management, error-detective, network-engineer, systematic-debugging, and incident-responder, using standard Linux tooling such as systemd/journalctl, ss, dig, strace, and lsof throughout.

SREs, sysadmins, and on-call engineers diagnosing a live Linux issue who want a phased checklist with the exact command and sub-skill for each diagnostic step, rather than improvising an investigation order.