Getting it into your agent
One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.
npx skills add mindspore-ai/akg --skill triton-ascend-api-rulesgit clone --depth 1 https://github.com/mindspore-ai/akgWrote 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.
[](https://agentmods.dev/skills/mindspore-ai/akg/triton-ascend-api-rules)<a href="https://agentmods.dev/skills/mindspore-ai/akg/triton-ascend-api-rules"><img src="https://agentmods.dev/badge/skills/mindspore-ai/akg/triton-ascend-api-rules/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.
<a href="https://agentmods.dev/skills/mindspore-ai/akg/triton-ascend-api-rules"><img src="https://agentmods.dev/badge/skills/mindspore-ai/akg/triton-ascend-api-rules.svg" alt="Reviewed on agentmods" width="80" height="20"></a>- NVIDIA SkillSpector pass
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.
| Model | Per session | Once invoked |
|---|---|---|
| Fable 5.1 | $0.00073 | $0.00600 |
| Opus 5 | $0.00036 | $0.00300 |
| Sonnet 5 | $0.00015 | $0.00120 |
| Haiku 4.5 | $0.00007 | $0.00060 |
Grade A, and why
triton-ascend-api-rules 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 9d ago.
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.
What it actually says
Triton Ascend API Hard Rules (MUST follow)
This file lists rules whose violation causes a compile or runtime error on Ascend. They apply to every kernel in this DSL — no exceptions.
禁止使用的语法
return/break/continue→ 使用 mask 控制- lambda → 内联函数或 tl.where
- 链式布尔运算 → 分步计算 mask
- 张量直接索引 → tl.load / tl.store
- if-else 中负偏移 → tl.maximum(offset, 0)
- Ascend: 复杂 tl.where → if-else
- Ascend: while 循环 → for 替代
- Ascend: range() 的 start/stop 混用运行时变量和 constexpr → 用全 constexpr 的 range + 循环体内运行时 if 跳过
While 循环替代（Ascend）
静态上限（编译时常量）: 直接 for i in range(N_ITERS)
动态上限（运行时参数）:
@triton.jit
def kernel(ptr, n_iters, TILE: tl.constexpr, MAX_ITERS: tl.constexpr):
for i in range(MAX_ITERS):
if i < n_iters:
offset = i * TILE + tl.arange(0, TILE)
data = tl.load(ptr + offset)
tl.store(ptr + offset, data * 2)
切片操作
- 禁止 Python 切片
b[0]b[i:j] - 单元素:
tl.get_element(tensor, (index,)) - 切片:
tl.extract_slice(tensor, offsets, sizes, strides) - 插入:
tl.insert_slice(full, sub, offsets, sizes, strides) - 禁止对 tl.arange 张量用 get_element
其他限制
- tl.constexpr 仅在内核参数中使用，host 侧不可用
- 输出张量用 torch.empty / empty_like（避免 zeros/ones 初始化开销）
- 标量转换仅
scalar.to(type)，禁止tl.float16(scalar) - BLOCK_SIZE 必须小于 65536
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.
- 9d ago First seen · 57 lines · 73 tokens per session scan A b5a02585c002
triton-ascend-api-rules is a skill published in the GitHub repository mindspore-ai/akg (259 stars, last pushed 1mo ago), licensed Apache-2.0. It adds 73 tokens to every session and 600 once invoked, about $0.0004 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 2026-09-03.
Other skills, from other repositories
spark-environment-setup
Set up a working ML training/inference environment on NVIDIA DGX Spark (GB10, aarch64, CUDA 13). Use when installing PyTorch/Unsloth/TRL/vLLM on DGX Spark, hitting libcudart or wheel-ABI errors on aarch64, or choosing between NGC containers and bare pip installs.
spark-memory-thermal-ops
Manage unified memory and thermals during long-running ML jobs on NVIDIA DGX Spark. Use when planning memory headroom for a training run on GB10, when a job OOMs on unified memory, or when monitoring temperature and power during multi-hour training.
spark-training-gotchas
Preflight and diagnose the ten known failure modes for ML training on NVIDIA DGX Spark. Use when a training run on DGX Spark fails to start, OOMs below the 128GB limit, slows down mid-run, or before any multi-hour training job on GB10.
llama-cpp
Runs LLM inference on CPU, Apple Silicon, and consumer GPUs without NVIDIA hardware. Use for edge deployment, M1/M2/M3 Macs, AMD/Intel GPUs, or when CUDA is unavailable. Supports GGUF quantization (1.5-8 bit) for reduced memory and 4-10× speedup vs PyTorch on CPU.
minicpm5-deploy-vllm-ascend
Deploy MiniCPM5-2B with vLLM on Huawei Ascend NPU using vLLM-Ascend. Use when the user mentions vLLM-Ascend, Ascend NPU, Huawei Ascend, CANN, torchnpu, davinci devices, or wants an OpenAI-compatible MiniCPM5 server on Ascend hardware.
minicpm5-deploy-litert
Run MiniCPM5-2B or MiniCPM5-1B on-device with Google's LiteRT-LM runtime — the litert-lm CLI or its OpenAI-compatible server on a desktop, the Kotlin API or the AI Edge Gallery app on Android, the same .litertlm bundle on CPU or GPU. Use when the user says "LiteRT", "LiteRT-LM", "litertlm", ".litertlm", "Android"…