| name | interpro-search |
| description | Identify protein domains, families, and sites; find proteins sharing a domain/family; explore species distribution; and get domain architectures via InterPro (integrating Pfam, CDD, and more). Public API, no credential needed. |
| license | Apache-2.0 |
Interpro Search (Codex-native)
Codex-native adaptation of Google DeepMind's interpro-database skill
(science-skills, Apache-2.0).
Uses the public API directly through Codex's own tools.
Gates (ask before proceeding)
- Terms notice: on first use in a workspace, tell the user to review the
source's terms of use and record the notice in provenance.
- Network / install: read-only public access; be polite about rate limits.
Write large responses to a file and parse them; if you write helper code, run
it via
uv run, never bare python3. Ask before installs or large downloads.
Endpoints
- InterPro API
https://www.ebi.ac.uk/interpro/api/
Workflow
Search InterPro entries first with science_search_interpro; use the direct API
for protein/structure/taxonomy joins and pagination.
- Look up an entry (
/entry/interpro/<IPR>) or the domains of a protein (/entry/all/protein/uniprot/<ACC>).
- Find members of a family or proteins with a domain via the entry->protein relationship endpoints.
- Provenance & review: record queries, endpoints, and cited identifiers with
$science-provenance; check claims with $science-review before presenting.
Boundaries
- Report the member database (Pfam, CDD, etc.) behind each InterPro entry.
- Cite only InterPro/member accessions actually returned.