kernel-dev-skill
- Repo stars 47
- Author updated Live
- Author repo MateBot
- Domain
- Other
- Compatible agents
-
- Claude Code
- Cursor
- Cline
- Codex
- Windsurf
- Gemini CLI
- +20
- Trust score
- 88 / 100 · community maintained
- Author / version / license
- @aresbit · no license declared
- Token usage
- Lean
- Setup complexity
- Guided setup
- External API key
- Not required
- Operating systems
- Linux
- Runtime requirements
- No special requirements
- Permissions
-
- Read-only
- Shell exec
- Write / modify
- Network behavior
- Local-only
- Install commands
- 26 variants
Profile is derived at build time from SKILL.md and install vectors. Subject to drift from author intent.
Heads up: 未限定 allowed-tools，默认拥有全部工具权限。
---
name: kernel-dev-skill
description: Linux kernel development skill grounded in local references/labs and references/lectures materia…
category: other
runtime: no special runtime
---
# kernel-dev-skill output preview
## PART A: Task fit
- Use case: Linux kernel development skill grounded in local references/labs and references/lectures materials. Use for kernel modules, system calls, process scheduling, interrupts, locking, memory management, filesystems, networking, architecture, debugging, profiling, and device model work..
- Inputs: target material, constraints, expected output, and acceptance criteria.
- Evidence boundary: follow “Use this skill when / Scope / Coverage contract” and do not present inference as author intent.
## PART B: Execution result
- **01** The card summarizes the use case; runtime output centers on “Linux kernel development skill grounded in local references/labs and references/lectures materials. Use for kernel modules, system calls, process scheduling, interrupts, locking, memory management, filesystems, networking, architecture, debugging, profiling, and device model work.”.
- **02** When the source has headings, the agent prioritizes “Use this skill when / Scope / Coverage contract” so the result follows the author’s structure.
- **03** Typical output includes task judgment, concrete steps, required commands or file edits, validation, and follow-up options.
- **04** Risk context follows the fingerprint: read files, run shell commands, write/modify files; mostly runs locally; usually needs no extra API key.
## Running Rules
- read files, run shell commands, write/modify files; mostly runs locally; usually needs no extra API key.
- Validate with a small sample before expanding scope.
- Return the result, validation criteria, and next iteration options. Choose the trigger
The source mentions slash commands such as `/lib`; use them first when your agent supports command triggers.
Provide inputs and boundaries
Name target files or source material, expected output, forbidden changes, and whether network or shell access is allowed. Permission fingerprint: read files, run shell commands, write/modify files.
Validate with a small run
Start with a small task and check whether the result follows “Use this skill when / Scope / Coverage contract”. Inspect diffs, logs, previews, or tests before expanding scope.
Review before delivery
Confirm the final output includes a concrete result, evidence, and next action. If it stays generic, tighten inputs, boundaries, and acceptance criteria.
---
name: kernel-dev-skill
description: Linux kernel development skill grounded in local references/labs and references/lectures materia…
category: other
source: aresbit/MateBot
---
# kernel-dev-skill
## When to use
- Linux kernel development skill grounded in local references/labs and references/lectures materials. Use for kernel mod…
- Use it when the task has clear inputs, repeatable steps, and validation criteria.
## What to provide
- Target material, scope, expected result, and forbidden changes.
- Whether network, commands, file writes, or external services are allowed.
## Execution rules
- Organize steps around “Use this skill when / Scope / Coverage contract” and keep inference separate from source facts.
- read files, run shell commands, write/modify files; mostly runs locally; usually needs no extra API key.
- Validate with a small sample before expanding the task.
## Output requirements
- Return the deliverable, key evidence, validation method, and next action.
- Mark missing information as unknown; do not invent commands, platforms, or dependencies. Evidence boundary and execution chain
The author source anchors workflow facts; repository files anchor sources and commands; Fluxly only adds fit, limitations, and quality judgment.
skill "kernel-dev-skill" {
input -> user goal + target files + boundaries + acceptance criteria
context -> Use this skill when / Scope / Coverage contract
rules -> SKILL.md triggers / order / output contract
runtime -> no special runtime | read files, run shell commands, write/modify files | mostly runs locally
guardrails -> usually needs no extra API key + small-sample validation + diff/log review
output -> copyable result + checklist + next iteration
}Linux Kernel Development
Core rule: simple is superior to everything.
This skill is for practical kernel engineering. Use it to analyze failures, choose the smallest relevant subsystem material, make minimal verifiable changes, and validate with evidence.
Use this skill when
- Kernel code fails to build or run on a target kernel
- A module, subsystem patch, or lab skeleton needs implementation
- You need help on process, syscall, interrupt, SMP, memory, filesystem, networking, or architecture topics
- You are decoding oops/panic, lock bugs, memory bugs, or performance regressions
- You are working on device-model or driver-adjacent kernel paths
Scope
Includes:
- kernel modules and build flow
- kernel API and execution context rules
- system calls and process interactions
- interrupts and deferred work
- SMP and synchronization
- memory management and mapping
- filesystems and VFS-facing logic
- networking stack and net path basics
- architecture layer and portability concerns
- debugging and profiling
- device and driver model
Coverage contract
When responding, ensure the chosen path explicitly maps to one of these kernel components:
- module lifecycle
- syscall boundary
- process/scheduler path
- interrupts/deferred work
- locking/SMP behavior
- memory subsystem (allocation, mapping, lifetime, reclaim-facing assumptions)
- filesystem/VFS path
- networking path
- architecture portability
- debugging/profiling or device model
If the issue touches memory, call out which memory aspect is involved:
- allocation/lifetime (
kmalloc,vmalloc,kzalloc, free path symmetry) - user/kernel copy boundary
- mapping or virtual memory behavior (
mmap, vm area assumptions) - context safety (sleeping/atomic constraints)
Excludes:
- pure user-space programs
- generic Linux administration without kernel code
Required inputs
Collect minimal hard evidence before proposing a patch:
- target kernel version, distro, architecture
- source path and build entry (
Makefile/Kbuild/ target) - first failing build log or runtime log
- subsystem guess: module, syscall, process, irq, memory, fs, net, arch, driver, or unknown
- desired outcome: compile fix, runtime fix, behavior change, or learning implementation
Workflow
- Freeze baseline. Run one build or collect one complete runtime failure trace.
- Classify failure. Place blockers into one class: API break, context violation, concurrency, memory, lifecycle, functional bug, or performance issue.
- Route to one reference first.
Use references/source-map.md to pick the smallest matching file from
references/labsorreferences/lectures. - Patch one theme per step. Avoid mixing unrelated refactors.
- Verify immediately. Rebuild and run one focused check that proves the specific blocker moved.
- Expand only after proof. After the first fix is validated, handle next blocker.
Hard rules
- Minimal, reversible, evidence-based changes.
- No API claims without code/log evidence.
- Do not redesign architecture while baseline is broken.
- Do not hide uncertainty; label assumptions and provide next confirming command.
- Keep recommendations subsystem-specific, not generic Linux advice.
Command set
make -C /lib/modules/$(uname -r)/build M=$PWD V=1 modules
make -C /lib/modules/$(uname -r)/build M=$PWD W=1 C=1 modules
dmesg -T | tail -n 200
scripts/checkpatch.pl --strict -f path/to/file.cAdd subsystem commands only when they directly test the current blocker.
Subsystem verification rule
For the selected subsystem, include at least one verification step with an expected signal:
- memory: allocator path, user-copy return handling, or mapping behavior check
- interrupts: handler registration and interrupt-path signal in logs
- process/scheduler: task state transition or wake/sleep behavior evidence
- filesystem: VFS callback path evidence
- networking: packet path or interface state transition evidence
- syscall boundary: errno and copy boundary behavior evidence
Output format
1. Subsystem and failure class
2. Evidence used
3. Relevant local material
4. Smallest patch sequence
5. Verification step and expected signal
6. Risks / assumptionsOutput must satisfy all items:
- name one primary kernel component from the coverage contract
- cite the local material file used for that component (
references/labs/*.mdorreferences/lectures/*.md) - provide one concrete verification step with expected signal
- if subsystem is unknown, state the shortest command to disambiguate it
Fast mode
For firefighting requests:
- show top 3 blockers only
- map each blocker to one local reference
- provide smallest next patch and one verification step
Decide Fit First
Design Intent
How To Use It
Boundaries And Review