astropy

Trigger phrases

Phrases that activate this skill when typed to Claude Code:

• FITS file
• celestial coordinates
• cosmological calculation
• astronomical data

What it does

astropy is a Claude Code skill from K-Dense AI’s scientific-agent-skills repo. It turns Claude into a working astronomer’s assistant — handling coordinate transformations, unit conversions, FITS file I/O, cosmological distance calculations, time scale conversions, and astronomical table operations using the Astropy library.

The output of a session is correct, unit-aware Python code. Astropy’s Quantity system propagates physical units through calculations, so the skill generates code that avoids unit mismatch bugs common in hand-rolled astrophysics scripts.

When to use it

Reach for it when:

• You need to read or write FITS files, extract image data, or parse telescope headers
• You are converting between coordinate systems (ICRS, Galactic, AltAz) or computing angular separations and sky crossmatches
• You need cosmological calculations (luminosity distance, comoving volume) or time scale conversions (UTC, TAI, TDB)

When not to reach for it:

• You need specialized radio astronomy or interferometry tooling (reach for CASA or related packages)
• Your task is entirely statistical and has no astronomical-specific data types involved

Install

Copy the SKILL.md from scientific-skills/astropy into .claude/skills/astropy/.

The skill activates on trigger phrases including “FITS file”, “celestial coordinates”, and “astronomical data”.

What a session looks like

A typical session has three phases:

1. Data ingestion. Claude opens a FITS file with astropy.io.fits, reads the primary HDU or table extension, and surfaces the header keywords relevant to the task (WCS, exposure time, filter, telescope).
2. Computation. Claude constructs SkyCoord objects, applies coordinate frame transforms, computes separations or crossmatches, or evaluates a cosmology model (e.g., Planck18) for distance calculations — all with unit tracking via astropy.units.
3. Output. Claude serializes results to an astropy.table.Table (CSV, FITS, or ECSV format) and generates a summary of derived quantities with their units explicitly stated.

Receipts

Honest reporting on what astropy handles well and where it falls short:

Where it works well:

• Crossmatching two sky catalogs by angular separation — SkyCoord.match_to_catalog_sky is concise and correct
• FITS header parsing for surveys with non-standard keyword conventions — Claude can map keywords to the right WCS parameters
• Cosmological distance calculations where getting units right manually is error-prone

Where it backfires:

• Very large FITS images (multi-GB mosaics) require memory-mapped access; the skill generates correct code but doesn’t always pre-empt RAM limit warnings
• Complex WCS with SIP distortions may need reproject or astropy-wcs-utils beyond what the base astropy skill covers

Pattern that works: always attach astropy.units to raw numeric inputs at the point of entry — catching unit errors there is far cheaper than debugging downstream calculation failures.

Source and attribution

Originally authored by K-Dense, Inc.. The canonical SKILL.md lives in the astropy folder of the scientific-agent-skills repository.

License: MIT. Install, adapt, and redistribute with attribution preserved.

This page documents the skill from a practitioner’s perspective. For the formal spec and updates, defer to the source repo.