onescience-resource-fetch SKILL.md skillA
onescience-resource-fetch from onescience-ai/oneskills is an agent skill: a SKILL.md folder an AI agent loads when a task calls for it.
mdr add onescience-ai/oneskills/onescience-resource-fetch@git:20261103.19e6633Pins git:20261103.19e6633 in mdr.lock. Pin the fingerprint instead.
- Audit grade
- A, 17 of 17 checks passed
- Kind
- SKILL.md skill
- Repository
- onescience-ai/oneskills
- Path
skills/onescience-resource-fetch/SKILL.md- Versions recorded
- 2
- Latest version
- git:20261103.19e6633, committed Nov 3, 2026
- First version recorded
- Nov 3, 2026
- Size
- 21,662 bytes
- Repository stars
- 24
- License
- MIT
OneSkills 资源层「物化」技能（type=resource），专管数据准备：任务需要的模型权重/checkpoint 或数据集在本地不存在时，优先复用共享目录 / 本地已有副本，缺失时从共享目录 / ModelScope / HuggingFace / 官网直链四类来源里，按 references/source_selection.md 评估择一，把资源拉到本地统一缓存并返回可用的绝对路径（ModelScope 是默认但非唯一、也非无条件最优的来源）。解析顺序固定四段式——任务级资源规划（请求里没有具体资源名时，先按 references/resource_planning.md 把任务拆成资源需求并派生检索词，不回问用户）→ 调用方显式路径 → 本地探测（~/.onescience 缓存、工作区 data/、注册表记录）→ 共享目录按名 find（可选，命中即免下载）→ 段3 网络物化；缺即下、不询问确认。段3 由调用方按 references/source_selection.md 在 ModelScope（默认）/ HuggingFace（--source hf…
Collected from public GitHub. Every version is saved exactly as it was and never changes. Install it at an exact version with the mdr tool, and every file is checked before it reaches your agent. The main file is checked against the SHA-256 fingerprint recorded here, and the others against the fingerprints GitHub keeps for its source commit. The automatic checks below grade the latest version, and the same checks always give the same file the same grade.
What the file says
# OneScience Resource Fetch 你是 OneSkills 资源层的物化技能（`type=resource`）。资源层的知识召回由 `onescience-primitives` 负责，**把资源真正取到本地**由你负责：输入一个逻辑资源名（数据集名 / 模型名），输出一个可直接使用的本地绝对路径 + 完整性结论，或如实报告取不到及其原因。 渐进披露：正文只保留触发、硬门禁与契约字段；下载细节、契约扩展、校验细则下沉 `references/`，在标注「必须先读取」的位置必须先读再做。 ## 何时使用 - 任务所需的**数据集**或**模型权重/checkpoint** 在本地、工作区、缓存里都不存在，需从共享目录 / ModelScope / HuggingFace / 官网直链择一取得（按 `references/source_selection.md` 评估，ModelScope 默认但非唯一最优）。 - 上游（orchestrator / executor）在 `blocked_reason=data_missing`、`model_checkpoint_missing`、`resource_acquisition` 类缺口上要求补数据准备。 - 用户直说"把 X 数据集下到本地""取一下 MACE/ERA5 的权重""准备训练数据"。 - 用户**只给了裸资源名**（没有组织名、没有 `repo_id`，甚至语种或大小写与仓库名不完全一致）：站内检索 + 打分自动择优后直接物化，不回问"完整仓库名是什么"。 - 用户**只给了任务描述**（连数据集/模型名都没提，如"训练一个猫狗图像分类器""复现这篇论文的 DFT 结果"）：先做资源规划再检索定位，把选中的仓库与理由一并交回。 ## 何时不要用 - 缺的是**知识**（论文、参数惯例、晶格结构）→ `onescience-primitives` / `onescience-online-knowledge`。 - 缺的是**软件包或运行环境**（pip/conda 库、VASP/LAMMPS）→ `onescience-installer`。 - 已有本地路径且只需**格式转换 / AI-Ready 标准化 / 切分** → `onescience-data-standardizer`。 - 需要**发布**资源到 ModelScope（上传）→ `onescience-modelscope-publish`。 - 用户显式禁止联网或要求只做只读校测 → 用 `probe` / `--probe-only`，不得下载。 ## 强制协议 …
Read the whole file at its exact version.
How to install
mdr add onescience-ai/oneskills/onescience-resource-fetch@sha256:7a3c84f0768928e9The line at the top of this page pins the label git:20261103.19e6633, which is what most people want. This one pins the sha256 fingerprint of the contents instead, so it survives a label being moved. Either way the exact version is written to mdr.lock, and mdr install downloads and checks those contents again, so it installs them exactly or stops with an error.
[](https://markdownregistry.com/a/art_fsvg6ekam2twlxcx)
0 badge image fetches in 30 days. What the badge shows.
Versions
The registry has recorded 2 versions of this file, all committed on Nov 3, 2026.
| version | committed | commit | size | audit | |
|---|---|---|---|---|---|
| git:20261103.19e6633 latest | 2026-11-03 | 19e6633 | 21,662 B | A | view · diff |
| git:20261103.1456029 | 2026-11-03 | 1456029 | 18,301 B | A | view |
Audit of the latest version
- pass: Frontmatter block present
- pass: Frontmatter declares a name
- pass: Frontmatter declares a description
- pass: Size between 200 bytes and 200 KB (21662 bytes)
- pass: No zero-width or bidi control characters
- pass: No instruction hidden inside an HTML comment
- pass: No link to an exfiltration or paste host
- pass: No credential-shaped string
- pass: No instruction to send local credentials anywhere
- pass: No text hidden with inline styles
- pass: No prompt-injection phrasing
- pass: No curl or wget piped into a shell
- pass: No recursive delete of root, home or parent
- pass: No instruction to read or print local credentials
- pass: No base64 blob over 200 characters
- pass: No link to a raw IP address
- pass: No script tag
Source
onescience-ai/oneskills · 24 stars · license MIT · pushed 2026-11-04 · branch master
API
GET https://markdownregistry.com/api/v1/artifacts/art_fsvg6ekam2twlxcx GET https://markdownregistry.com/api/v1/resolve?ref=onescience-ai/oneskills/onescience-resource-fetch GET https://markdownregistry.com/api/v1/blob/7a3c84f0768928e9637f94ee006d6d15325c4a638bc3367cdaf390826c9b8325
Your AI assistant does the legwork. You hear about the deals worth your word. Hand yours the standing instructions at modelranch.com and it joins the network that reads files like this one.
Guides: the SKILL.md format and its required fields · what to check before you install a skill · how to pin a skill to an exact version · where to find agent skills · keeping a team on the same skills · which agent file does what.