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 sparse_quantgit 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/sparse_quant)<a href="https://agentmods.dev/skills/mindspore-ai/akg/sparse_quant"><img src="https://agentmods.dev/badge/skills/mindspore-ai/akg/sparse_quant/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/sparse_quant"><img src="https://agentmods.dev/badge/skills/mindspore-ai/akg/sparse_quant.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 | $0.00132 | $0.03270 |
| Opus 5 | $0.00066 | $0.01635 |
| Sonnet 5 | $0.00026 | $0.00654 |
| Haiku 4.5 | $0.00013 | $0.00327 |
Grade A, and why
sparse-quantization 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.
How it starts
The opening of the file, as written. The whole thing — 285 lines — stays where its author put it; the contents beside it link to each section on GitHub.
稀疏量化模型加载与适配 Skill
背景知识
vLLM-MindSpore 基于 vLLM 二次开发适配华为昇腾 NPU，稀疏量化（W8A8SC）是一种在 310P 推理机上支持的量化方案，使用 Golden Stick 量化工具链生成量化模型。
实现依据：vllm_mindspore PR !1428（W8A8SC 基础设施）、!1490（310P/910 兼容、权重加载优化、测试样例）
| Rule | Value |
|---|---|
| 量化格式 | W8A8SC（weight 8bit, activation 8bit, sparse compressed） |
| 支持推理机 | 310P（910 不支持） |
| 核心文件 | vllm_mindspore/.../golden_stick/ |
| 量化工具 | Golden Stick（msModelSlim） |
设计方案：vLLM-MindSpore 官方稀疏量化设计文档 → Native 模型支持稀疏量化设计方案与代码改动说明
加载流程架构
文件路径 | 作用 | 关键函数
vllm_mindspore/.../golden_stick/golden_stick.py | 量化方法注册与识别 | get_quant_method(), get_config_filenames()
vllm_mindspore/.../golden_stick/a8w8sc.py | A8W8SC 线性层实现 | A8W8SCLinearMethod.create_weights()
vllm_mindspore/.../quant_ops.py | 稀疏量化算子封装 | QuantLinearSparseOp
vllm_mindspore/.../models/sparse_quant_weight_loader.py | 权重加载与 310P 格式转换 | load_split_weights(), _param_name_to_weight_key()
vllm_mindspore/.../sparse_quant_loader.py | 稀疏量化模型加载器 | SparseQuantModelLoader
vllm_mindspore/config.py | 稀疏量化配置识别 | is_sparse_quantization
vllm_mindspore/.../qwen2.py | 模型权重加载 | load_weights()
加载流程详解
模型目录扫描
│
▼
┌─────────────────────────────────────────┐
│ 1. 配置文件识别 (golden_stick.py) │
│ - quantization_description.json │
│ - quant_model_description.json │
│ - quant_model_description_w8a8sc.json│
└─────────────────────────────────────────┘
│
▼
┌─────────────────────────────────────────┐
│ 2. 稀疏量化判断 (golden_stick.py) │
│ - 配置中存在 rank_ 开头的键 │
│ - 设置 load_format=sparse_quant │
└─────────────────────────────────────────┘
│
▼
┌─────────────────────────────────────────┐
│ 3. 推理机校验 (a8w8sc.py) │
│ - 310P: 支持稀疏量化 │
│ - 910: 报错 INFERENCE_910_SPARSE_QUANT│
└─────────────────────────────────────────┘
│
▼
┌─────────────────────────────────────────┐
│ 4. 权重加载 (sparse_quant_weight_loader)│
│ - 从 rank_{tp_rank} 目录读取权重 │
│ - deq_scale: float32 → int64(310P) │
└─────────────────────────────────────────┘
│
▼
推理就绪
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 · 285 lines · 132 tokens per session scan A d0fc8ba21012
sparse-quantization is a skill published in the GitHub repository mindspore-ai/akg (259 stars, last pushed 1mo ago), licensed Apache-2.0. It adds 132 tokens to every session and 3,270 once invoked, about $0.0007 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"…