The Agent Skills Directory

[PROMPT_INJECTION]: The skill is designed to ingest and process data from external astronomical files, creating a surface for indirect prompt injection.\n
Ingestion points: Functions like Table.read(), fits.open(), and CCDData.read() in SKILL.md and references/data_io_guide.md load data from external formats such as FITS, CSV, and VOTable.\n
Boundary markers: There are no explicit instructions or delimiters defined to prevent the agent from interpreting ingested data content as instructions.\n
Capability inventory: The agent has the ability to write files to the local disk using writeto() and write() methods, as documented in SKILL.md and references/data_io_guide.md.\n
Sanitization: The skill relies on standard library parsing without additional sanitization or content validation for the ingested data.\n- [EXTERNAL_DOWNLOADS]: The skill documentation includes examples of fetching data from remote locations.\n
The astropy.utils.data.download_file function is documented in references/auxiliary_modules.md for fetching remote resources.\n
Support for remote FITS file access (e.g., via S3) is mentioned in references/data_io_guide.md.

astropy-astronomy