AstroDB Bot

Agent Skills for interactive AstroDB ingestion workflows. Expects an input data table and will prepare it to be converted into a new database using the AstroDB Template schema and conventions. It utilizes the Felis system for describing database schemas. Also uses astrodbkit and astrodb_utils packages for setting up and interacting with the database.

More detailed instructions on how to use these skills are in the AstroDB Toolkit documentation [Creating a Database with AI Skills] (https://astrodb-utils.readthedocs.io/en/stable/pages/db_management/agent_skills.html).

To install this in another agent, you can copy the skills/ directory to whatever is appropriate for your agent (eg, .claude/skills/, .cursor/skills/, .agents/skills/, etc).

Skills

Build

• astrodb-build-01-setup — Set up the local directory, cloning the astrodb template and naming it properly.
• astrodb-build-02-parse-table — Summarize table columns, descriptions, units, and types.
• astrodb-build-03-schema-match — Map parsed columns to AstroDB Template tables and fields.
• astrodb-build-04-schema-validate — Identify problems with nulls and inconsistent data types.
• astrodb-build-05-schema-generate — Create a Felis-format schema.yaml file using outputs of previous skills.
• astrodb-build-06-create-db — Create an empty SQLite AstroDB database from a Felis-validated schema.yaml.
• astrodb-build-07-erd — Generate an entity relationship diagram from a Felis schema.yaml.

Ingest

• astrodb-ingest-01-publications — Ingest publications into the Publications table from a DOI/bibcode or a table of references; backfills missing metadata for existing references.
• astrodb-ingest-02-sources — Ingest astronomical objects into the Sources table from a data file.
• astrodb-ingest-03-photometry — Ingest magnitudes into the Photometry table from a data file, setting up the PhotometryFilters/Telescopes/Instruments (from SVO filter IDs) first.

Website

• astrodb-website — Set up a FastAPI web interface (astrodb-web) to browse and visualize an AstroDB SQLite database.

Requirements

• an AI skill runner that reads .agents/skills/ (or the appropriate directory for your agent)
• uv to install Python packages
• Python 3.13+
  • astropy
  • pandas for fallback table parsing
  • lsst-felis
  • astrodbkit
  • astrodb_utils
  • pytest

License

BSD 3-Clause. See LICENSE.