Liger Kernel Dev
Developmentliger-kernel-devLiger Kernel Dev is an Agent Skill for development published in the linkedin/Liger-Kernel repository on GitHub: a SKILL.md instruction file an AI agent can load to do this task.
Develops production-ready Triton kernels for Liger Kernel. Creates new kernels from PyTorch operations (local files, URLs, code snippets, or natural language) with ops, module wrappers, functional APIs, unit tests, benchmarks, and plots. Also modifies existing Liger kernels. Use when adding a new Triton kernel, converting a PyTorch operation to Triton, or updating an existing Liger kernel.
How to use this skill
Bring this guide into your coding agent with a prompt tailored to the tool you use.
- Open your project in Codex.
- Copy and paste the prompt below.
- Review files and risks before approving any action.
Inspect the Agent Skill at https://github.com/linkedin/Liger-Kernel/blob/HEAD/.agents/skills/liger-kernel-dev/SKILL.md and all supporting files. Report provenance, permissions, scripts, dependencies, and security risks. Propose an exact installation into .agents/skills/liger-kernel-dev, but wait for my approval before writing files or running scripts.
Copying this prompt does not install or run the skill. Review third-party files before use.
Liger Kernel Dev
Develops Triton kernels for Liger Kernel through a 3-stage pipeline with human review between stages. Supports creating new kernels and modifying existing ones. NVIDIA GPUs only.
Mode Detection
- Create mode: User asks to create/add/generate/write/build a new kernel → full pipeline
- Modify mode: User asks to update/fix/change/extend an existing kernel → skip Analyze, modify files, then Validate
Pipeline (Create Mode)
Stage 1: Analyze
Follow the Analyzer workflow in analyzer.md. If the host runtime supports parallel subagents, this stage may be delegated to one; otherwise execute the workflow directly.
Accepts any input: local file, URL, code snippet, natural language description, or model component reference. Produces a standalone PyTorch reference implementation and a kernel profile.
Human checkpoint: Present PyTorch reference + kernel profile. Confirm before proceeding.
Stage 2: Generate
Follow the Generator workflow in generator.md.
Generates/modifies up to 8 files:
src/liger_kernel/ops/{kernel}.py— NEW Triton kernels + autograd Functionsrc/liger_kernel/transformers/{kernel}.py— NEW nn.Module wrappersrc/liger_kernel/transformers/functional.py— MODIFY add functional APIsrc/liger_kernel/ops/__init__.py— MODIFY export Function classsrc/liger_kernel/transformers/__init__.py— MODIFY export Module +__all__test/transformers/test_{kernel}.py— NEW unit testsbenchmark/scripts/benchmark_{kernel}.py— NEW benchmark scriptbenchmark/data/all_benchmark_data.csv— MODIFY (after benchmarks run)
Human checkpoint: Present changes for review.
Stage 3: Validate
Follow the Validator workflow in validator.md.
Runs checkstyle, unit tests (hard gate — stops on persistent failure), benchmarks, and generates plots. Optionally runs ncu profiling.
Human checkpoint: Report final results with benchmark numbers and plots.
Pipeline (Modify Mode)
- Read existing kernel files to understand current implementation
- Understand the requested modification
- Make targeted changes (Generator handles this)
- Run full Validate stage (same as create mode)
Reference Files
- kernel-profile-format.md — Kernel profile schema and field descriptions
- examples/swiglu-profile.md — Tier 1 (element-wise) reference
- examples/rms-norm-profile.md — Tier 2 (reduction) reference
- examples/cross-entropy-profile.md — Tier 3 (fused/complex) reference
- Templates in templates/ — Code generation patterns for each file type