claudemods
中文English

Modsfiftyone-dataset-curation

fiftyone-dataset-curation e28a9a3d453b Immutable snapshotLatest

Pinned to commit e28a9a3d453b15cbefa6ccc81c89fecf723c31b2, captured .

Download e28a9a3d453b source archive Choose something to compare with View commit on GitHub

More actions and subscriptions

All versions

Current package description (not historical)

End-to-end dataset curation for any FiftyOne dataset: inspect schema and media quality (blur, brightness, corruption, resolution), audit annotation quality (mistakenness, hardness, duplicate labels, IoU overlaps), analyze class distributions and imbalances, explore embedding space and data gaps, find near/exact duplicates, create curated subsets using semantic search and brain ops, build train/val/test splits, and answer natural language questions about your data. Delegates to fiftyone-find-duplicates and fiftyone-embeddings-visualization. Requires FiftyOne MCP server with @voxel51/brain and @voxel51/utils plugins.

Install

This version is listed in the verified marketplace fiftyone-skills (repository voxel51/fiftyone-skills). Run these in Claude Code, in order:

  1. Add the marketplace/plugin marketplace add voxel51/fiftyone-skills
  2. Install the plugin/plugin install fiftyone-dataset-curation@fiftyone-skills

These commands install the marketplace's current upstream state, not the pinned commit snapshot on this page; the two can differ.

Collection warnings
  • no plugin.json at the package root; components inferred from the marketplace entry and default directories
  • author not declared; showing repository owner

Contents 1

Skill1
Dependencies, compatibility and skill declarations

Recorded dependencies 0

No dependency requirements were recorded from plugin.json or marketplace entries. Only manifest fields saved with the snapshot are shown here; dependency declarations in other files are outside the current collection scope.

Recorded compatibility

No compatibility requirement was recorded for this snapshot from plugin.json or a marketplace entry. Only manifest fields saved with the snapshot are shown here. Declarations in a skill's SKILL.md header are collected separately and, when present, shown on their own under “Skill file declarations”; they never overwrite or fill in this value.

Skill file declarations 1

The content below was read separately from the header of each skill file (SKILL.md) at its pinned commit. It is text the file declares about itself, shown as an annotation only. It is stored apart from the snapshot's dependencies and compatibility (manifest fields) above and does not overwrite them. It is not the package license, version, install count or a release record, and none of it is verified.

Evidence scope and limits
  • Each entry is one file at one pinned commit, and field states hold for that file only. Components that point to the same file are merged into one entry; files with identical content share one extraction.
  • “Declared in file” means the file's header has the field. “Not declared in this file” appears only when the header parsed successfully and lacks the field. “Invalid declaration” means the field is present but its type, length or characters do not meet the requirements; it is never rewritten or truncated. “Not read” means no usable header was obtained (not yet read, read failed, no header, or an unparseable header), which says neither that the field was declared nor that it was not.
  • Compatibility and license text are the file's own wording. They do not show that anything runs, are not legal advice, and are not a license verification. metadata.version is only an annotation inside the file; it is not the package version and is not used to judge which is newer.
  • This list groups the version's skill components by their recorded source. A listed source may still be unpinned, unread or unavailable; each entry shows its state. Collection runs in bounded batches and does not promise complete coverage.
  • A file whose read failed shows a “retry eligible” time. It is only when the collector may try again, not a schedule, and it does not promise a read or completion.

1–1 of 1 skill file sources · As of

  1. skills/fiftyone-dataset-curation/SKILL.md

    Not read

    Components using this file 1

    • fiftyone-dataset-curation skills/fiftyone-dataset-curation/SKILL.md

    Fields declared in the file

    Compatibility (declared in file)
    Not readNo usable header, which says neither that it was declared nor that it was not.
    License text (declared in file, unverified)
    Not readNo usable header, which says neither that it was declared nor that it was not.
    metadata.version (file annotation)
    Not readNo usable header, which says neither that it was declared nor that it was not.

    This file has no usable extraction record yet, so all three fields show as “Not read”.

    Repository
    voxel51/fiftyone-skills
    Pinned commit
    e28a9a3d453b15cbefa6ccc81c89fecf723c31b2
    File identity
    Regular-file identity recorded
    Blob SHA
    8da0d795d5f2d098426758a73abe98ad29f7cb7e
    Mode
    100644
    Identity recorded
    Header parse
    Not read
    Why not read
    The file identity is recorded, but its content has not been read yet. declaration_unread

    No extraction record yet.

Read this page as JSON

Versions of this package1
VersionCommitCapturedComponentsActions
e28a9a3d453b Latest e28a9a3 1 View Download Compare…