UI はローカライズ済みです。スキルの紹介と説明文書は選択した言語へ AI が自動翻訳し、コード・コマンド・パス・専門用語は原文のままです。原作者の文書が優先されます。
スキルを探す
データ・研究 · SKILL.md

datasets-permissions

datasets-permissions

Operate PySyft dataset publication, mock/private resolution, syft URLs, and SyftBox permission APIs.

Independent GitHub snapshot · Automated classification · Original documentation governs

スキルの説明

GitHub

name: datasets-permissions description: "Operate PySyft dataset publication, mock/private resolution, syft URLs, and SyftBox permission APIs." disable-model-invocation: true metadata: disco-role: operating license: Apache 2.0

datasets-permissions

Use this sub-skill for create_dataset, mock/private dataset separation, resolve_dataset_file_path, load_dataset_code, syft:// paths, dataset visibility/deletion, upload_private, syft.pub.yaml, syft_permissions, and syft_perms ACL questions.

Workflow

  1. Confirm DO and DS are accepted peers and synced; otherwise route to ../auth-sync-transport/SKILL.md.
  2. DO creates a dataset with public mock files and private owner-side files.
  3. DS lists/reads mock files after sync and tests analysis locally.
  4. Inside a submitted job, resolve_dataset_file_path(...) resolves the owner-side private file. Outside a job, use client= for mock testing.
  5. Validate permission rules and explain nearest-syft.pub.yaml behavior before changing ACLs.

Read references/dataset-workflows.md, references/permissions-reference.md, references/data-formats.md, references/api-reference.md, and references/troubleshooting.md.

Helpers: scripts/create_dataset_fixture.py creates tiny mock/private fixtures; scripts/check_permission_yaml.py checks a simple local ACL rule.

保存したスキル