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 op-task-extractorgit 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/op-task-extractor)<a href="https://agentmods.dev/skills/mindspore-ai/akg/op-task-extractor"><img src="https://agentmods.dev/badge/skills/mindspore-ai/akg/op-task-extractor/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/op-task-extractor"><img src="https://agentmods.dev/badge/skills/mindspore-ai/akg/op-task-extractor.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 | $0.00027 | $0.01158 |
| Opus 5 | $0.00014 | $0.00579 |
| Sonnet 5 | $0.00005 | $0.00232 |
| Haiku 4.5 | $0.00003 | $0.00116 |
Grade A, and why
op-task-extractor 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 — 122 lines — stays where its author put it; the contents beside it link to each section on GitHub.
What I do
将用户的输入(代码文件或自然语言描述)转化为标准格式的 {op_name}.py 任务文件,并通过验证脚本确认可执行。
When to use me
需要为后续算子生成工作流准备标准化任务文件时
Workflow
Step 1 判断输入类型
├─ 用户提供了符合格式要求的代码文件 → 保存为 {op_name}.py → Step 4(验证)
│ ├─ 通过 → 返回(任务结束)
│ └─ 失败 → Step 2(分析并重写)
└─ 用户提供自然语言描述 → Step 2
Step 2 代码分析 & 依赖追踪
Step 3 构建 {op_name}.py(参考末尾「输出格式」)
Step 4 运行验证脚本(必须执行)
Step 5 用户确认
Step 1: 判断输入类型
- 用户提供了符合格式要求的代码文件 → 直接保存为
{op_name}.py,跳到 Step 4 运行验证脚本 - 其他 → 进入 Step 2
Step 2: 代码分析 & 依赖追踪
- 读取用户确认的
framework,backend,arch配置 - 如有源代码:
- 识别待优化部分,提取 shape/dtype 信息,确定输入/输出签名
- 分析依赖关系(AST 级别),追踪自定义函数/类,确定需要内联的外部依赖
- 如为自然语言描述:
- 从描述中理解算子语义,确定合理的 shape/dtype 默认值
Step 3: 构建 {op_name}.py
按末尾「输出格式」生成文件,用 PyTorch/Python 实现:
- 将算子逻辑包装到
Model.forward()中 - 如有初始化状态(权重、参数),放入
Model.__init__() - 将所有自定义依赖内联到文件中
- 根据 shape/dtype 信息构建
get_inputs()和get_init_inputs() - 如用户未提供 shape/dtype,从代码上下文推断合理默认值
Step 4: 运行验证脚本(必须执行)
使用命令模板执行 @scripts/validate_kernelbench_task.py:
python <本skill绝对路径>/scripts/validate_kernelbench_task.py \
/abs/path/{op_name}.py --json
验证脚本同时执行静态检查(4 个组件齐全)和运行时检查(实例化、前向传播、NaN/Inf、一致性)。
结果处理:
- 输出
[VALID]+ 来源是 Step 1 的用户原始文件 → 直接返回,任务结束 - 输出
[VALID]+ 来源是 Step 3 新生成的文件 → 进入 Step 5 - 输出
[INVALID]→ 根据错误信息修复代码,重新验证(最多 2 次) - 重试 3 次仍失败 → 向用户报告错误,请求协助
Step 5: 用户确认
任务描述文件内容非用户提供的原始代码时,必须执行
将完整的 {op_name}.py 内容展示给用户,使用 question 工具请求确认。
不通过则结合用户反馈返回 Step 3。
关键约束
| 约束 | 说明 |
|---|---|
| 自包含 | 所有依赖函数必须内联,禁止 import 项目内模块 |
| 可执行 | Model(*get_init_inputs()).forward(*get_inputs()) 必须直接运行 |
| 确定性 | 给定相同输入,输出必须一致 |
| 无 NaN/Inf | forward 输出不能包含 NaN 或 Inf |
| 禁止重写 | 原始函数可运行就直接复用,一行都不改 |
| 返回一致 | 返回类型/形状必须与原始实现一致 |
| 合理输入 | get_inputs 应提供合理大小的输入(不能过小或过大) |
输出格式
最终文件必须是单一自包含 Python 文件,包含以下 4 个部分:
# 1. Imports 区(只允许标准库和 PyTorch 相关包)
import torch
import torch.nn as nn
# 2. Model 类
class Model(nn.Module):
def __init__(self, <init_params>):
super(Model, self).__init__()
def forward(self, <forward_inputs>) -> torch.Tensor:
return output
# 3. get_inputs():返回 forward() 的输入参数列表
def get_inputs():
input1 = torch.randn(batch_size, dim)
input2 = torch.randn(batch_size, dim)
return [input1, input2]
# 4. get_init_inputs():返回 __init__() 的初始化参数列表
def get_init_inputs():
return [dim_value]
What ships with it
1 file 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.
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 · 122 lines · 27 tokens per session scan A ccb3abc0b08f
op-task-extractor is a skill published in the GitHub repository mindspore-ai/akg (259 stars, last pushed 1mo ago), licensed Apache-2.0. It adds 27 tokens to every session and 1,158 once invoked, about $0.0001 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
matlab
Build, review, migrate, and safely plan MATLAB or GNU Octave numerical workflows, including arrays, tabular/time data, tests, projects, graphics, MAT files, and explicit Python interoperability.
pennylane
Hardware-agnostic quantum ML framework with automatic differentiation. Use when training quantum circuits via gradients, building hybrid quantum-classical models, or needing device portability across IBM/Google/Rigetti/IonQ. Best for variational algorithms (VQE, QAOA), quantum neural networks, and integration with…
dd-code-generation
Use pup CLI for immediate Datadog operations or generate code for integration into applications.
rocm-kernels
Provides guidance for writing and benchmarking optimized Triton kernels for AMD GPUs (MI355X, R9700) on ROCm, targeting HuggingFace diffusers (LTX-Video, SD3, FLUX) and transformers. Core kernels: RMSNorm, RoPE 3D, GEGLU, AdaLN. Includes XCD swizzle, autotune, diffusers integration patterns, and LTX-Video pipeline…
holoscan-install-wheel
Install Holoscan SDK Python wheel via pip into a venv. Use for Python installs; not for native C++/apt or Conda installs.
typing-exclusion-worker
Python typing exclusion worker: remove assigned mypy exclusion modules in small scoped batches, fix typing issues, run validation, and produce a structured completion summary. Use when running parallel typing-debt workers or when asked to remove modules from pyproject mypy exclusion overrides.