Reproducibility

Every tool in the BioContext7 index carries provenance metadata that traces its origin from upstream registries through compilation to the API response. This page documents how to verify and reproduce any snapshot.

Verify a tool snapshot

Tool Provenance Lookup

Search for any tool to see its upstream version, generation timestamp, source hash, and compiler version. Each field links back to the source commit and crawl log entry.

Use the tool search to look up a specific tool. Every result includes:

upstream_version — resolved from bioconda, PyPI, or GitHub Releases
generated_at — ISO 8601 compilation timestamp
source_hash — SHA-256 of all compilation inputs (click to expand)
generator_version — skill compiler semantic version
Link to raw SKILL.md on GitHub at the exact commit
Link to the matching crawl-log.jsonl entry

Snapshot statistics

Aggregate provenance metrics from tools-stats.json:

1,800+

Total skills

88%

Verified versions

12%

Unknown versions

Weekly

Refresh cadence

Schemas

tools-index.schema.json

Defines the compact tool index array. Rejects any record with v == "latest" via a JSON Schema "not" constraint.

skills-index.schema.json

Declares required provenance fields (tool_id, version, upstream_version, generated_at, generator_version, source_hash) on every entry.

tools-provenance.schema.json

Defines the ProvenanceRecord shape with upstream, snapshot, and compiler sub-objects for external validation.

Re-derive this snapshot

git clone --depth 1 https://github.com/Hordago-Labs/biocontext7
cd biocontext7 && git checkout v0.5.0
python packages/biocontext7/scripts/recompile-skills.py \
    --from-log packages/biocontext7/data/crawl-log.jsonl \
    --output ./reproduced-skills
diff -r ./reproduced-skills plugins/biocontext7-skills/skills/

The --from-log flag replays the frozen crawl log instead of hitting live APIs, ensuring byte-for-byte reproducibility.

Citation

@software{biocontext7,
  author    = {Jaureguy, Jeff},
  title     = {BioContext7: Bioinformatics Registry Aggregator},
  version   = {0.5.0},
  note      = {Zenodo DOI pending — will be minted on the first public GitHub release},
  url       = {https://github.com/Hordago-Labs/biocontext7},
  date      = {2026-04-06},
}

The data_version and generated_at fields are interpolated from the current tagged release. The Zenodo DOI will appear here once the first public release is archived — set NEXT_PUBLIC_ZENODO_DOI to publish it.