技能 fiftyone-dataset-import
📦
fiftyone-dataset-import
內容修訂 r1 高風險 ⚙️ 外部命令🌐 網路存取📁 檔案系統存取
將視覺資料集匯入 FiftyOne
當媒體、標註和多模態資料夾使用不同格式時，資料集匯入常會失敗。此技能會引導你在 FiftyOne 中進行掃描、偵測、套件設定、匯入、分組和驗證。
支援: Claude Codex Code(CC)
安裝前請停止並請求確認。
請審查安裝計畫，並在變更檔案前取得使用者明確同意。
使用我的智慧代理安裝
將此請求複製給你的智慧代理，其中包含權威 Skill 頁面和清單。
智能體請求
Review the Skillstore skill "fiftyone-dataset-import" from https://skillstore.io/skills/adonaivera-fiftyone-dataset-import.md and its manifest at https://skillstore.io/api/skills/adonaivera-fiftyone-dataset-import/manifest. Verify the artifact. Stop and obtain explicit user consent before installing or changing files.你的智慧代理仍應顯示其計畫，並依安全政策請求必要確認。
Agent 可讀資源
當 AI Agent、爬蟲或腳本需要乾淨脈絡、而不是讀取完整頁面時，請使用這些連結。
測試它
正在使用「fiftyone-dataset-import」。 從我的本機資料夾匯入 COCO 影像資料集。
預期結果:
- 掃描摘要列出影像數量、標註檔案和偵測到的 COCO 結構。
- 建議的匯入設定包括資料集名稱、標籤欄位和來源路徑。
- 助理會在建立資料集前請求確認。
正在使用「fiftyone-dataset-import」。 載入包含相機和 LiDAR 的自動駕駛資料集。
預期結果:
- 偵測到的切片包括前視相機、側視相機和 LiDAR。
- 列出所需套件和點雲轉換步驟。
- 驗證會比較預期群組、樣本和標籤欄位。
正在使用「fiftyone-dataset-import」。 排查失敗的 FiftyOne 匯入問題。
預期結果:
- 回應會檢查路徑、支援的副檔名、標籤格式和既有資料集名稱。
- 它會建議先進行較小規模的測試匯入，再重試完整資料集。
安全審計
高風險v8 • 2026/7/5 開啟版本化報告
Most static hits are Markdown backticks, code fences, examples, or benign FiftyOne references. Confirmed issues are limited to package installation guidance and Python import checks that can execute third-party code. A high-risk semantic issue remains because the skill recommends pickle.load on dataset files.
1
已掃描檔案
1,089
分析行數
18
複核項目
0
已忽略的誤報
已確認的安全問題 (1)
高
Unsafe Pickle Loading From Dataset Files
The skill shows pickle.load on annotation and LiDAR files from dataset paths. Malicious pickle files can execute code during loading.
The cited examples directly load pickle data from dataset files. Python pickle is unsafe for untrusted input and can execute code when deserialized.
能力複核項目 (18)
這些是真實的本機能力，對此技能而言可能是預期行為，因此需要複核，但不會按已確認的惡意行為計分。
中
Ruby/shell backtick execution
| PandaSet | `pandaset` | `pip install "git+https://github.com/scaleapi/pandaset-devkit.git#subdirec
The skill recommends pip installation of third-party packages. Package installation can execute setup code and modify the user environment.
中
Ruby/shell backtick execution
| nuScenes | `nuscenes-devkit` | `pip install nuscenes-devkit` |
The skill recommends pip installation of third-party packages. Package installation can execute setup code and modify the user environment.
中
Ruby/shell backtick execution
| Argoverse 2 | `av2` | `pip install av2` |
The skill recommends pip installation of third-party packages. Package installation can execute setup code and modify the user environment.
中
Ruby/shell backtick execution
| KITTI 3D | `pykitti` | `pip install pykitti` |
The skill recommends pip installation of third-party packages. Package installation can execute setup code and modify the user environment.
中
Ruby/shell backtick execution
| Lyft L5 | `l5kit` | `pip install l5kit` |
The skill recommends pip installation of third-party packages. Package installation can execute setup code and modify the user environment.
中
Ruby/shell backtick execution
| Point cloud conversion to PCD | `open3d` | `pip install open3d` |
The skill recommends pip installation of third-party packages. Package installation can execute setup code and modify the user environment.
中
Ruby/shell backtick execution
| Point cloud processing | `pyntcloud` | `pip install pyntcloud` |
The skill recommends pip installation of third-party packages. Package installation can execute setup code and modify the user environment.
中
Ruby/shell backtick execution
| LAS/LAZ point clouds | `laspy` | `pip install laspy` |
The skill recommends pip installation of third-party packages. Package installation can execute setup code and modify the user environment.
中
Ruby/shell backtick execution
2. **Test import** in Python: `python -c "from <package> import ..."`
The skill recommends running a Python import check for a newly installed package. Importing untrusted package code can execute local code as a side effect.
中
Ruby/shell backtick execution
**Note:** Install open3d if needed: `pip install open3d`
The skill recommends pip installation of third-party packages. Package installation can execute setup code and modify the user environment.
低
Hardcoded URL
| PandaSet | `pandaset` | `pip install "git+https://github.com/scaleapi/pandaset-devkit.git#subdirec
The URL is used for remote package installation or cloning. Fetching and installing code from external repositories introduces supply-chain risk.
低
Hardcoded URL
pip install "git+https://github.com/<org>/<repo>.git"
The URL is used for remote package installation or cloning. Fetching and installing code from external repositories introduces supply-chain risk.
低
Hardcoded URL
pip install "git+https://github.com/<org>/<repo>.git#subdirectory=python"
The URL is used for remote package installation or cloning. Fetching and installing code from external repositories introduces supply-chain risk.
低
Hardcoded URL
pip install "git+https://github.com/<org>/<repo>.git@v1.0.0"
The URL is used for remote package installation or cloning. Fetching and installing code from external repositories introduces supply-chain risk.
低
Hardcoded URL
git clone https://github.com/<org>/<repo>.git
The URL is used for remote package installation or cloning. Fetching and installing code from external repositories introduces supply-chain risk.
低
Hardcoded URL
pip install "git+https://github.com/scaleapi/pandaset-devkit.git#subdirectory=python"
The URL is used for remote package installation or cloning. Fetching and installing code from external repositories introduces supply-chain risk.
低
Hardcoded URL
pip install "git+https://github.com/<org>/<repo>.git#subdirectory=python"
The URL is used for remote package installation or cloning. Fetching and installing code from external repositories introduces supply-chain risk.
低
Hardcoded URL
git clone https://github.com/<org>/<repo>.git && cd <repo> && pip install .
The URL is used for remote package installation or cloning. Fetching and installing code from external repositories introduces supply-chain risk.
風險因素
⚙️ 外部命令 (155)
SKILL.md:27 SKILL.md:28 SKILL.md:36-40 SKILL.md:40-48 SKILL.md:48 SKILL.md:55-65 SKILL.md:65 SKILL.md:66 SKILL.md:67-68 SKILL.md:68 SKILL.md:69 SKILL.md:70 SKILL.md:71-77 SKILL.md:77 SKILL.md:78 SKILL.md:79 SKILL.md:84-86 SKILL.md:86-89 SKILL.md:89-98 SKILL.md:98-101 SKILL.md:101-105 SKILL.md:105-110 SKILL.md:110 SKILL.md:111 SKILL.md:117-118 SKILL.md:118-131 SKILL.md:131-133 SKILL.md:133-136 SKILL.md:136-151 SKILL.md:151-163 SKILL.md:163-172 SKILL.md:172-178 SKILL.md:178-186 SKILL.md:186 SKILL.md:187 SKILL.md:188 SKILL.md:189 SKILL.md:197 SKILL.md:198 SKILL.md:199 SKILL.md:200 SKILL.md:201 SKILL.md:202 SKILL.md:203 SKILL.md:204-205 SKILL.md:205 SKILL.md:206 SKILL.md:207 SKILL.md:208 SKILL.md:214 SKILL.md:215 SKILL.md:216 SKILL.md:217 SKILL.md:218 SKILL.md:219 SKILL.md:225-230 SKILL.md:230-237 SKILL.md:237-243 SKILL.md:243-244 SKILL.md:244-253 SKILL.md:253-256 SKILL.md:256-265 SKILL.md:265-268 SKILL.md:268-270 SKILL.md:270-273 SKILL.md:273-275 SKILL.md:275-288 SKILL.md:288-297 SKILL.md:297-301 SKILL.md:301-309 SKILL.md:309-313 SKILL.md:313-318 SKILL.md:318-325 SKILL.md:325-360 SKILL.md:360-372 SKILL.md:372-375 SKILL.md:375-381 SKILL.md:381 SKILL.md:385-392 SKILL.md:392-396 SKILL.md:396-408 SKILL.md:408-414 SKILL.md:414-436 SKILL.md:436-442 SKILL.md:442-488 SKILL.md:488-497 SKILL.md:497-498 SKILL.md:498-504 SKILL.md:504 SKILL.md:506-531 SKILL.md:531-533 SKILL.md:533-537 SKILL.md:537-539 SKILL.md:539-555 SKILL.md:555-561 SKILL.md:561-565 SKILL.md:565-571 SKILL.md:571 SKILL.md:572 SKILL.md:573 SKILL.md:574 SKILL.md:575 SKILL.md:576 SKILL.md:580-690 SKILL.md:690-698 SKILL.md:698-703 SKILL.md:703-710 SKILL.md:710-721 SKILL.md:721-725 SKILL.md:725-731 SKILL.md:731-740 SKILL.md:740-745 SKILL.md:745-753 SKILL.md:753 SKILL.md:754 SKILL.md:755 SKILL.md:756 SKILL.md:762 SKILL.md:763 SKILL.md:764 SKILL.md:765 SKILL.md:766 SKILL.md:767 SKILL.md:768-769 SKILL.md:769 SKILL.md:770 SKILL.md:771-772 SKILL.md:772-773 SKILL.md:773-774 SKILL.md:774-775 SKILL.md:775-776 SKILL.md:776 SKILL.md:777 SKILL.md:778 SKILL.md:779 SKILL.md:780-786 SKILL.md:786-809 SKILL.md:809-813 SKILL.md:813-835 SKILL.md:835-839 SKILL.md:839-861 SKILL.md:861-867 SKILL.md:867-914 SKILL.md:914-918 SKILL.md:918-940 SKILL.md:940-944 SKILL.md:944-979 SKILL.md:979-988 SKILL.md:988-989 SKILL.md:989-991 SKILL.md:991-1006 SKILL.md:1006-1020 SKILL.md:1020-1035 SKILL.md:1035-1043 SKILL.md:174
🌐 網路存取 (13)
📁 檔案系統存取 (1)
審計單位: codex 檢視審計紀錄 →
分享與引用此報告
分享版本化評估報告、中性徽章、嵌入卡片與引用。Skillstore 只報告證據，不判斷該 Skill 是否安全。
複製報告連結
https://skillstore.io/skills/adonaivera-fiftyone-dataset-import/audits/8?utm_source=security_passport&utm_medium=share&utm_campaign=versioned_reportMarkdown 徽章
[](https://skillstore.io/skills/adonaivera-fiftyone-dataset-import?utm_source=security_passport_badge)HTML 徽章
<a href="https://skillstore.io/skills/adonaivera-fiftyone-dataset-import?utm_source=security_passport_badge"><img src="https://skillstore.io/badges/skills/adonaivera-fiftyone-dataset-import/security.svg" alt="Skillstore security assessment" loading="lazy"></a>嵌入卡片
<iframe src="https://skillstore.io/embed/skills/adonaivera-fiftyone-dataset-import.html" title="Skillstore Security Assessment" sandbox="allow-popups allow-popups-to-escape-sandbox" loading="lazy" referrerpolicy="no-referrer" width="420" height="180"></iframe>學術引用 (APA · BibTeX · CFF)
APA 引用
AdonaiVera. (2026). fiftyone-dataset-import security audit report (audit version 8) [Author version unspecified]. Skillstore. https://skillstore.io/skills/adonaivera-fiftyone-dataset-import/audits/8BibTeX 引用
@techreport{adonaivera-adonaivera-fiftyone-dataset-import-2026,
author = {AdonaiVera},
title = {fiftyone-dataset-import security audit report (audit version 8)},
institution = {Skillstore},
year = {2026},
number = {8},
url = {https://skillstore.io/skills/adonaivera-fiftyone-dataset-import/audits/8},
note = {Author version unspecified}
}CITATION.cff
cff-version: 1.2.0
message: "If you use this Skill, cite its author and this versioned security audit report."
title: "fiftyone-dataset-import security audit report (audit version 8)"
version: "unspecified"
type: report
authors:
- name: "AdonaiVera"
date-released: "2026-07-05"
url: "https://skillstore.io/skills/adonaivera-fiftyone-dataset-import/audits/8"
identifiers:
- type: other
value: "skillstore:adonaivera-fiftyone-dataset-import:audit:8"
description: "Skillstore immutable audit report identifier"
Skillstore 評分
評分依據 證據置信度: 中 55
架構
85
可維護性
87
內容
73
社群
70
規範符合度
你可以打造什麼
匯入標準偵測資料集
將帶有 COCO、YOLO、VOC、CVAT 或 KITTI 標籤的影像資料集載入新的 FiftyOne 資料集。
建立多模態駕駛資料集
依場景分組相機影像、LiDAR 點雲和標籤，以供自動駕駛審查使用。
驗證資料集遷移結果
在分析前，將匯入的樣本數、標籤、群組和切片與來源檔案進行比較。
試試這些提示
匯入簡單影像資料集
掃描我位於 [path] 的影像資料集，偵測標籤，提出 FiftyOne 匯入計畫，並等待我的確認。
偵測未知資料集格式
檢查 [path]，識別媒體類型和標註格式，然後建議最安全的 FiftyOne 匯入設定。
建立多模態群組
分析 [path] 中的相機、LiDAR 和場景分組模式。提出群組切片和驗證檢查清單。
規劃自訂自動駕駛匯入
為位於 [path] 的 [dataset name] 建立 FiftyOne 匯入計畫，包括所需套件、轉換步驟和標籤對應。
最佳實務
- 在建立資料集前掃描資料夾並呈現發現結果。
- 在安裝套件或覆寫資料集前請求確認。
- 每次匯入後驗證樣本數、標籤和群組。
避免
- 不要在與使用者確認偵測到的格式前進行匯入。
- 不要將應保持分組的多模態場景攤平成單一結構。
- 不要在沒有安全審查流程的情況下載入不可信任的 pickle 檔案。
常見問題
此技能需要 FiftyOne 嗎？
是。它假設 FiftyOne、FiftyOne MCP server 和列出的 Voxel51 plugins 均可用。
它可以匯入 YOLO 和 COCO 資料集嗎？
可以。此技能包含 YOLO、COCO、VOC、CVAT、KITTI 和其他格式的偵測與匯入指引。
它可以處理點雲嗎？
可以。它涵蓋 PCD、PLY、LAS、LAZ，並提供其他點雲格式的轉換指引。
它會自動安裝套件嗎？
指示要求在安裝套件前取得使用者確認。核准前請審查套件來源。
它可以覆寫既有資料集嗎？
它會檢查既有資料集名稱，並詢問是否覆寫、重新命名或中止。
它對不可信任的標註檔案安全嗎？
請謹慎使用。此技能包含 pickle 範例，而 pickle 檔案在載入時可能執行程式碼。
開發者詳細資訊
作者
AdonaiVera授權
MIT
Skillstore 修訂
r1
版本提示
作者未聲明版本。
引用
6eec4a59f63ff83440a84c18d1aef79743bc714e
維護新鮮度
2026/7/22
使用量
6 次下載 · 378 次瀏覽
檔案結構
📄 SKILL.md