vllm-ascend-workspace: Skill for Codex

.agents/skills/ascend-triton-operator-development/SKILL.md

ascend-triton-operator-development is a skill for Codex from maoxx241/vllm-ascend-workspace. It costs 118 tokens per session (648 once invoked), scanned A, original, MIT.

A workflow for writing or moving Triton GPU kernels to Ascend processors. Triton is a language for GPU and accelerator code; Ascend is Huawei's accelerator hardware.

In plain words
What is it for?
Auditing kernel behavior, designing hardware-aware grids and tiles, implementing the kernel, and handing it to remote Ascend validation.
Why use it?
It turns a PyTorch reference or existing kernel into a documented candidate and requires correctness checks before claiming it works.

Skill for Codex

Written for Codex: agents/openai.yaml present. Also seen: installed under .agents/ (shared by several agents).

This is maoxx241/vllm-ascend-workspace's own configuration. It tells Codex how to work on vllm-ascend-workspace itself, so it is not a mod to install elsewhere. Copy it as a starting point and replace the rules that are about this project. Everything vllm-ascend-workspace configures →

Reuse

Borrowing it

Nothing to install: this file belongs to maoxx241/vllm-ascend-workspace. Take a copy, put it at the same path in your own repository, and replace the rules that are about this project with yours.

Copy the file
curl -O https://raw.githubusercontent.com/maoxx241/vllm-ascend-workspace/main/.agents/skills/ascend-triton-operator-development/SKILL.md
Clone the repo
git clone --depth 1 https://github.com/maoxx241/vllm-ascend-workspace

Made for: Codex.

Wrote this? Show the measurements

A badge with what this costs and how it scanned, read live from this page, so it follows the numbers instead of freezing them. Markdown for a README, HTML for a documentation site or a project page.

agentmods badge for ascend-triton-operator-development

README.md
[![agentmods](https://agentmods.dev/badge/skills/maoxx241/vllm-ascend-workspace/ascend-triton-operator-development/github.svg)](https://agentmods.dev/skills/maoxx241/vllm-ascend-workspace/ascend-triton-operator-development)
Your own site
<a href="https://agentmods.dev/skills/maoxx241/vllm-ascend-workspace/ascend-triton-operator-development"><img src="https://agentmods.dev/badge/skills/maoxx241/vllm-ascend-workspace/ascend-triton-operator-development/github.svg" alt="Measured on agentmods" height="20"></a>

Or the 80×15 button, for a site that already has a row of RSS and ATOM ones. Only the verdict fits; the numbers stay here.

agentmods 80×15 button for ascend-triton-operator-development

Your own site · 80×15
<a href="https://agentmods.dev/skills/maoxx241/vllm-ascend-workspace/ascend-triton-operator-development"><img src="https://agentmods.dev/badge/skills/maoxx241/vllm-ascend-workspace/ascend-triton-operator-development.svg" alt="Reviewed on agentmods" width="80" height="20"></a>
Per session 118 Skills are progressive disclosure: only the name and description are preloaded; the body loads when the skill is used.
When invoked 648 The whole file, excluding the scripts and references it only reads on demand.
Security scan A 0 findings. A grade says what 26 rules found in the file — not that it is safe. Third-party audits
  • NVIDIA SkillSpector pass 7 Sept 2026
How audits are shown
Origin original No closer match found in the catalogue.
Token cost

What it costs to keep this loaded

Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.

ModelPer sessionOnce invoked
Fable 5.1 $0.00118 $0.00648
Opus 5 $0.00059 $0.00324
Sonnet 5 $0.00024 $0.00130
Haiku 4.5 $0.00012 $0.00065

Measured 12d ago against content hash e125803f09f1, method: parsed. Prices are Anthropic first-party input rates as of 2026-09-11, from the pricing page.

Security

Grade A, and why

ascend-triton-operator-development scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 12d ago.

The scan reads SKILL.md. This mod also ships 2 executable files (scripts/triton_development.py, tests/test_triton_development.py), listed below but not scanned — reading those needs a real analyzer, not pattern matching.

A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.

Nothing flagged

None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.

.agents/skills/ascend-triton-operator-development/SKILL.md · 44 lines

How it starts

The opening of the file, as written. The whole thing — 44 lines — stays where its author put it; the contents beside it link to each section on GitHub.

Ascend Triton Operator Development

Produce a traceable candidate and prove its correctness through the validation Skill before claiming success.

Workflow

  1. Record the exact source, reference, target SoC, CANN/Triton-Ascend versions, supported shapes, dtypes, layouts, strides, scalar options, tolerances, and side effects.
  2. Query .agents/knowledge/ for target capability and known failure signatures. Treat absent facts as unknown.
  3. Run scripts/triton_development.py plan to create the task contract and development Run Manifest.
  4. Complete the generated semantic report before changing code. For GPU Triton input, audit every load, store, mask, index, grid dimension, reduction identity, atomic, and alias.
  5. Write one hardware-aware sketch: logical work, physical-core mapping, tile sizes, estimated UB live set, padding semantics, and specialization boundaries.
  6. Implement the smallest correct candidate. Keep the host wrapper limited to allocation, metadata extraction, dispatch, and launch; keep core computation in @triton.jit.
  7. Use ascend-triton-kernel-validation on a managed remote NPU. Do not run torch_npu locally.
  8. Run finalize with the candidate, completed audit, sketch, and terminal validation manifest.

Entry point

scripts/triton_development.py provides:

  • plan: validate the task config and create audit/sketch templates plus Run Manifest v1;
  • finalize: hash and register the candidate artifacts, consume terminal correctness evidence, and generate the development report.

Read:

Rules

  • Preserve the reference semantics; do not optimize away masks, padding, dtype width, or side effects without proof.
  • Do not hard-code core count, UB capacity, alignment, or compiler capability across SoCs and versions.
  • Keep all planned cases; never reduce a multi-shape task to the easiest case.
  • Do not use GPU latency as the NPU acceptance baseline.
  • A development run passes only when its linked validation manifest passes.
  • Keep run state under .vaws-local/ascend-triton/development/.

Read the full file on GitHub · 44 lines

Files

What ships with it

8 files beside SKILL.md in the same directory: the scripts, references and assets a skill reads on demand. Not counted in the per-session cost; read them before you install if any of them is executable.

Changes

What this file has done since we first saw it

Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.

  1. 12d ago First seen · 44 lines · 118 tokens per session scan A e125803f09f1

Subscribe to this mod's changes

ascend-triton-operator-development is a skill published in the GitHub repository maoxx241/vllm-ascend-workspace (36 stars, last pushed 7d ago), licensed MIT. It adds 118 tokens to every session and 648 once invoked, about $0.0006 per session on Opus 5. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-30.

Related

Other skills, from other repositories

gke-compute-classes

Configures, optimizes, and troubleshoots GKE ComputeClasses. Use when configuring Spot VMs with on-demand fallback, targeting specific accelerators (GPUs/TPUs) or machine families, restricting ComputeClass access, or debugging pending pods related to node pool auto-creation. Do not use for cluster-level Node Auto…

google/skills · 83 tokens

jetson-diagnostic

Read-only Jetson health snapshot for identity, memory, GPU, thermal, power, storage, services, and top processes.

NVIDIA/skills · 30 tokens

doca-socket-relay

Use this skill when the operator is driving the DOCA Socket Relay to bridge a socket-oriented host application onto a BlueField DPU peer without rewriting it — picking the deployment shape (in-process, sidecar, or BlueField service container), configuring the host-side socket and the DPU-side forwarding endpoint…

NVIDIA/skills · 236 tokens

offensive-z-wave

Z-Wave attack methodology — sniffing with Z-Force / EZ-Wave / RTL-SDR + ZniffMobile, S0 (legacy) network-key derivation flaw and key reuse, S2 (modern) ECDH commissioning analysis, replay/injection on unauthenticated nodes, default-key brute-force on test deployments, and home-automation hub pivots. Use when targeting…

SnailSploit/Claude-Red · 113 tokens

hsb-flash

Flash the FPGA on an HSB board connected to an NVIDIA devkit. Supports HSB Lattice boards (FPGA versions 2407, 2412, 2507, 2510) and Leopard Imaging VB1940 "all-in-one" cameras (FPGA versions 2507, 2510). Uses release-specific YAML manifests and board-type-specific program commands. Lattice and VB1940 commands must…

NVIDIA/skills · 94 tokens

jetson-validate-image

Use after jetson-flash-image to run static BSP checks, on-target smoke/regression tests on a flashed DUT, or both. Not for build or flash steps. Triggers: validate bsp, on-target validation.

NVIDIA/skills · 50 tokens