Generate verified Croissant metadata files for machine-learning datasets from crawled URLs.
npx skills add https://github.com/cbizon/claussant --skill croissant
Please help me install this Agent Skill. Skill: croissant Source: https://github.com/cbizon/claussant/tree/main/.claude/skills/croissant Command: npx skills add https://github.com/cbizon/claussant --skill croissant
This Skill automates the creation of machine-readable dataset descriptors (Croissant files) for ML datasets, ensuring all metadata is traceable to crawled sources and not the model's training knowledge.
_croissant.json, _provenance.json, and _incomplete.json files for datasets.Use the croissant skill to generate a Croissant file for the dataset at https://www.bindingdb.org.
High-intent search queries and answers about installing and using this skill.
To generate verified Croissant metadata, provide a publicly accessible dataset URL. The Skill crawls the source to extract identity, licensing, files, and schema, outputting a Croissant 1.1 compliant JSON file.
Croissant metadata provides a machine-readable dataset descriptor for ML frameworks and catalogues. It structures dataset identity, licensing, and schema information into a standardized, verifiable format.
Yes, generating a Croissant file requires a public dataset URL. The Skill enforces strict verification rules by crawling the URL, ensuring metadata is traceable to authoritative sources rather than training knowledge.
To ensure metadata is traceable to crawled sources, this Skill enforces a strict "no training knowledge" rule. It extracts all licensing, identity, and schema information directly from the provided public URL.
Creating a Croissant dataset descriptor generates three files: `_croissant.json` for the metadata, `_provenance.json` for traceability, and `_incomplete.json` for missing or unverified dataset elements.
This Skill generates new Croissant 1.1 specification files by crawling public URLs. It does not validate pre-existing files, but produces verified metadata packages that adhere to the MLCommons standard.
samuelmukotiAutomates web data collection via CLI and Python scripts with HTTP fetching, rendering, anti-bot bypass, and spider crawling.
GonglitianValidate dataset formats, schema consistency, and data quality for ML pipelines.
HolobiomicsLabParse FCS files and extract event data as NumPy arrays or CSV.
shekerkammaConnect CSV, DuckDB, MotherDuck, PostgreSQL, BigQuery, and Snowflake data sources.