Linux Kernel Development

What it does

Provides targeted, evidence-driven guidance for fixing and testing Linux kernel code. It helps the agent classify kernel failures, pick the smallest relevant reference, propose minimal patches, and verify results with concrete commands and expected signals. Typical outputs include a short diagnosis, a minimal patch sequence, and a verification step with expected logs or behaviors.

When to use it

• When kernel modules fail to build or load.
• To triage oops/panic or lock/memory bugs.
• For small, reversible kernel patches or driver fixes.
• When you need subsystem-specific verification (memory, irq, syscall, fs, net).

What's included

• Scripts: verification/build commands shown in the skill body (e.g. make module build, dmesg checks).
• References: the skill includes references/ folders and a source-map to local lecture/lab materials.
• Instructions: a disciplined workflow for freezing a baseline, classifying the failure, routing to the smallest local reference, making minimal changes, and verifying immediately. Hard rules emphasize minimal, evidence-backed changes.

Compatible agents

Best used by code-capable assistants with shell access and build tooling (Copilot/Codex-like agents, Claude Code, or other code-capable LLMs) that can run build/test commands and read logs.