agentmods.dev

vllm-ascend-workspace: Skill for Codex

.agents/skills/ascend-triton-workflow/SKILL.md

ascend-triton-workflow is a skill for Codex from maoxx241/vllm-ascend-workspace. It costs 93 tokens per session (569 once invoked), scanned A, original, MIT.

A workflow for developing and improving Ascend Triton operators, which are GPU-style computing kernels adapted for Ascend NPU hardware. It coordinates development, correctness checks, profiling, optimization, and run reports across multiple stages.

In plain words
What is it for?
Use it for GPU-to-NPU migrations or new Ascend Triton kernels when the work includes at least two lifecycle stages or needs a complete workflow report.
Why use it?
It provides an ordered process for a kernel project that spans development, testing, and performance work. It also records which stages were completed, failed, or left untested.

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 →

not rated 36repo 7d ago A scan SkillSpector: pass 93 tokens original MIT
View source ↗ maoxx241/vllm-ascend-workspace
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-workflow/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-workflow

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

Your own site · 80×15
<a href="https://agentmods.dev/skills/maoxx241/vllm-ascend-workspace/ascend-triton-workflow"><img src="https://agentmods.dev/badge/skills/maoxx241/vllm-ascend-workspace/ascend-triton-workflow.svg" alt="Reviewed on agentmods" width="80" height="20"></a>
Per session 93 Skills are progressive disclosure: only the name and description are preloaded; the body loads when the skill is used.
When invoked 569 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 $0.00093 $0.00569
Opus 5 $0.00046 $0.00284
Sonnet 5 $0.00019 $0.00114
Haiku 4.5 $0.00009 $0.00057

Measured 12d ago against content hash 0e3d06b86eb3, 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-workflow 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_workflow.py, tests/test_triton_workflow.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-workflow/SKILL.md · 45 lines

How it starts

The opening of the file, as written. The whole thing — 45 lines — stays where its author put it; the contents beside it link to each section on GitHub.

Ascend Triton Workflow

Coordinate the lifecycle without duplicating the implementation owned by the stage Skills.

Workflow

  1. Freeze the source, operator contract, target SoC, software versions, case set, and requested performance objective.
  2. Query .agents/knowledge/ for relevant capability, validation, and failure-signature facts. Treat missing facts as unknown.
  3. Run scripts/triton_workflow.py plan to create the stage plan and parent Run Manifest.
  4. Execute required stages with their owners:
    • first correct kernel or GPU migration: ascend-triton-operator-development;
    • compile and full-case correctness gate: ascend-triton-kernel-validation;
    • single-kernel profiling and performance iteration: ascend-triton-kernel-optimization.
  5. Before remote execution, establish remote-code-parity; run torch_npu and Triton only in a managed Ascend environment.
  6. Link every child Run Manifest to its planned stage with link.
  7. Run finalize and deliver the workflow report with missing, failed, and untested coverage explicit.

Entry point

scripts/triton_workflow.py provides:

  • plan: validate the workflow config, create ordered stage items, and initialize a parent Run Manifest;
  • link: bind one terminal child Run Manifest to one stage without overwriting prior evidence;
  • finalize: aggregate required-stage evidence and produce workflow-report.md.

Read:

Rules

  • Keep stage ownership strict; do not implement migration, validation, or optimization inside this Skill.
  • Never let performance evidence substitute for correctness evidence.
  • A passed workflow requires every required stage to have a passed terminal child manifest.
  • A failed child makes the workflow failed; missing, unsupported, cancelled, or inconclusive required evidence makes it inconclusive.
  • Record GPU timing only as cross-platform context; use a comparable NPU baseline for NPU optimization acceptance.
  • Keep orchestration state under .vaws-local/ascend-triton/workflows/.

Read the full file on GitHub · 45 lines

Files

What ships with it

6 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 · 45 lines · 93 tokens per session scan A 0e3d06b86eb3

Subscribe to this mod's changes

ascend-triton-workflow is a skill published in the GitHub repository maoxx241/vllm-ascend-workspace (36 stars, last pushed 7d ago), licensed MIT. It adds 93 tokens to every session and 569 once invoked, about $0.0005 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.

Hardware and IoTTesting
Related

Other skills, from other repositories

jetson-video-pipeline

Use when executing and verifying Jetson Video Codec SDK or PyNvVideoCodec encode/decode, transcode, segmentation, container decode, AV1, or acceptance workflows with exact artifact handoffs.

NVIDIA/skills · 45 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

holohub-app-lifecycle

Use for non-failing HoloHub app work with ./holohub: scaffold, build, run, test, visual evidence, lint, and flow benchmarking.

NVIDIA/skills · 40 tokens

code-plan

Turn a task description and repository into a structured implementation plan (files to create, files to modify, tests to add, risks).

initializ/forge · 29 tokens

ros2-robotics

Best practices for ROS 2 robotics development, covering package structure, nodes, topics/services/actions, launch files, QoS, tf2 transforms, and testing. Use when creating ROS 2 packages, writing nodes in rclpy or rclcpp, defining custom messages/services/actions, writing launch files, configuring QoS profiles…

Mindrally/skills · 87 tokens

SmartHome Video Anomaly Benchmark

VLM evaluation suite for video anomaly detection in smart home camera footage.

SharpAI/DeepCamera · 20 tokens