SKILLEMALL.ai
GitHub ↗ RU Sign in

AA kernel-cute-writing

Write and implement GPU kernels using NVIDIA CuTe DSL (CUTLASS 4.x Python API) — NOT for Triton, CUDA C++, or conceptual explanations. Trigger only when the user wants to write or implement a kernel, not when asking questions about CuTe DSL concepts or layouts. CuTe DSL uses cute.jit/cute.kernel decorators and cutlass.cute imports. Covers element-wise kernels, GEMM patterns, reductions, memory hierarchy (global/shared/register/TMA), MMA tensor core operations, software pipelining, and framework integration.

The skillemall take

Promises to write and run GPU kernels using CuTe DSL from CUTLASS 4.x—Python API for NVIDIA tensor operations. Triggers only on actual kernel implementation requests, not conceptual questions. Covers element-wise ops, GEMM, reductions, memory hierarchy (global/shared/registers/TMA), MMA with tensor cores, software pipelining, and framework integration.

Files check out: 20 files, 3 scripts, no critical issues. Scores solid (quality 93, safety 98), but process execution at 80 suggests some automation bottlenecks. Two medium-level findings exist but aren't blocking. Supports all major platforms.

Worth installing if you're writing CUDA-optimized code and comfortable with Python API instead of C++. Good for tensor operation experiments.

NVIDIA/TensorRT-LLM Agent Skills author: NVIDIA NOASSERTION 20 files · 3 scripts body ≈ 4 095 tokens Open the sourcegithub.com↗ analyzed 5 d ago

Write and implement GPU kernels using NVIDIA CuTe DSL (CUTLASS 4.x Python API) — NOT for Triton, CUDA C++, or conceptual explanations. Trigger only when the…

As a process A 80/100 · Runs to the end — no weak spots found

GeneratorSoftware developmenttype and topics are labelled automatically from the skill text
Runs in: Claude / Claude Code ChatGPT / Codex Gemini CLI Cursor GitHub Copilot / VS Code Qwen Code Kimi Code CLI Z.ai (GLM) DeepSeek (Deep Code) Mistral Vibe LangChain (deepagents) OpenClaw Goose OpenCode
JSON
Technical rating
A
96/100
safety, quality, tests
Safety 60%
98
Quality 40%
93
Run on models
none yet
Process rating
A
80/100
Runs to the end
Tools and files w 18
60
Result and completion w 14
60
Inputs and preconditions w 11
70
the three weakest of ten parameters · all ten

How to improve

    For the model run — optional
    • Your own cases (evals/evals.json, 4–6 real requests with expected answers): the full check would then run those instead of a model-drafted suite.
    • A spec.yaml with trigger phrases and assertions — a behaviour contract for CI; `skilltest init` writes a template.

    Guard findings · 2

    ✓ No critical or high findings

    Medium and low: 2
    • low Secrets in code secret-high-entropy-token references/patterns-memory.md:152
      High-entropy token-like string (may be an id, hash or a credential)
      Copy…tOp,  # Global → Shared (multicast)
    • low Secrets in code secret-high-entropy-token references/patterns-memory.md:153
      High-entropy token-like string (may be an id, hash or a credential)
      Copy…GOp,     # Shared → Global (with reduction)

    Files scanned: 20. Evidence is masked. Grey chips explain why severity was lowered.

    Against the Agent Skills spec

    ✓ No remarks against the Agent Skills spec

    Process rating: all ten parameters 80/100

    • 60Tools and files. Uses tools (web, python) that frontmatter does not declare
    • 60Result and completion. Output format stated, no completion criterion
    • 70Inputs and preconditions. Inputs and preconditions are listed
    • 70Execution cost. Instruction body is 4095 tokens
    • 85Steps. 68 steps, 1 vague phrases
    • 100When it triggers. States when to use and when not to
    • 100Failures and branches. 1 branches, has a failure section
    • 100Consistency. Name and required fields are in place
    • 100Running it twice. Mutating operations check current state
    • 100Progress reporting. Reports progress
    • medium Safety rules and hard prohibitions inside a skill: they belong in the system prompt, here they protect nothing

    Everything here is measured from the skill text rather than judged by a model, so the numbers are checkable. A parameter weighs more when it is a more common reason for the process to stall.

    Quality signals

    • +5Description has no quoted example phrases that should trigger the skill
    • -31 of 3 scripts are never mentioned in SKILL.md
    • +2Single-language instructions
    • +4Description says when NOT to use the skill
    • +3Description length 512: enough signal without eating the budget
    • +4Structure: 18 headings
    • +3Step-by-step instructions: 68 items
    • +3Output format is stated explicitly
    • +4Has examples (9 code blocks)
    • +4Reference files are cited in the instructions (16 of 16)
    • +1License stated

    Quality base 70; lint remarks subtract, signals add up to 100. Result: 93.