Skip to content
Agentmart
Submit a tool

dataset-curation

Curate generated images into a clean training dataset, validate metadata, remove duplicates, and sync manifests/assets to Hugging Face.

Skill by bl1nk42

IndexedApache-2.0★ 2
Write a review

Install

Works with Claude Code, Claude apps, Codex, Cursor, Gemini CLI, GitHub Copilot, VS Code, OpenCode, Goose, Windsurf. Read the repository before you give any tool access to your data.

npx agentmarket install skill/bl1nk42-dataset-curation

Asks which of your tools to add it to (Claude Code, Codex, Cursor, VS Code, Windsurf, Gemini CLI, Claude Desktop) and shows the change first. See https://agentmart.dev/cli.

Get the weekly diff

The weekly diff: one email a week with new agent tools, notable updates and grade changes, plus security alerts for tools you follow. Double opt-in, one-click unsubscribe.

We use your email only for this newsletter. Privacy

About

Curate generated images into a clean training dataset, validate metadata, remove duplicates, and sync manifests/assets to Hugging Face.

Use cases: ml

No reviews yet. Installed it? Reviews come from verified installers only. How reviews work

Permissions

What this skill can do and what it asks for, from its SKILL.md. Checked Oct 25, 2026.

Tools
No tools found
Secrets
None asked for
Install scripts
None found

Network hosts it talks to 1

  • codeload.github.com · found in install command

Files it touches

  • ~/.claude/skills/dataset-curation

What's inside

Skill name
dataset-curation
Description
Curate generated images into a clean training dataset, validate metadata, remove duplicates, and sync manifests/assets to Hugging Face.
License
Apache-2.0
Instructions
8 lines in SKILL.md · view on GitHub

Files (1)

  • SKILL.md

SKILL.md (first lines)

# Dataset Curation

Use this skill when the user asks to collect, clean, label, export, or publish image-generation training data. Never publish automatically. First inspect the manifest with `export_dataset_manifest`, validate that each row has an executed prompt, raw concept, image path, model, style, tags, and quality score, then check duplicates and negative votes.

Use `save_training_pair` for a local append-only record. Use `sync_hf_dataset` only after the user explicitly asks to sync or publish and the repository/token are configured. Prefer a dataset repository for CSV/JSONL and assets; use a model repository for model weights or LoRA files. Report repository URL and commit result.

Recommended lifecycle: collect -> validate -> deduplicate -> review negatives -> export manifest -> sync to Hugging Face.

From bl1nk42/image-gen, Apache-2.0.

More from bl1nk42/image-gen

Signals

Stars
2
Forks
0
Open issues
0
Last commit (this folder)
Sep 25, 2026 (1 month ago)

From GitHub (fetched Oct 24, 2026).

Details

License
Apache-2.0
First seen
Oct 24, 2026
Last checked
Oct 24, 2026

More from bl1nk42

Cite this page

dataset-curation skill. Agent Mart. https://agentmart.dev/skills/bl1nk42-dataset-curation (last updated 2026-10-24)