What problem does it solve? Analysis artifacts like downloaded files, processed TSVs, and model outputs often lose track of where they came from, making results impossible to reproduce or audit. This Skill records the sha256 hash, source URL, DOI, upstream version, download timestamp, and analysis parameters in a .meta.json sidecar next to every artifact. ## Core Features & Use Cases - Library and CLI modes: Import write_meta() into any Python script after saving output, or run stamp/verify/show subcommands from the command line on existing files. - Integrity verification: The verify command recomputes the artifact's sha256 and compares it against the sidecar, exiting non-zero on mismatch so corrupted or stale files are caught. - Cross-skill compatibility: The sidecar schema is a superset of the meta.json files produced by the gene-set-fetch, gene-annotations, gene-statistics, and ontology-terms skills, and verify() accepts both sha256 and output_sha256 fields. - Use Case: After downloading a GO annotation file and generating a filtered gene set TSV, stamp the output with the source URL, DOI, upstream release date, and filter parameters so a reviewer can reproduce the exact result months later. ## Quick Start Ask the AI to stamp an existing output file with its source URL, DOI, data version, and parameters, then verify the artifact's integrity against the generated sidecar.