The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: The skill makes network requests to rest.uniprot.org and www.uniprot.org to search for and retrieve protein information. These are official endpoints for a well-known and trusted scientific resource.
[COMMAND_EXECUTION]: Documentation files contain curl command examples for testing API endpoints. These are standard instructional examples for developers and do not involve executing external scripts or piping remote content to a shell.
[INDIRECT_PROMPT_INJECTION]: The skill ingests data from external API responses which theoretically provides a surface for indirect prompt injection. However, UniProt is a curated scientific database with low risk of adversarial content.
Ingestion points: User-provided search queries and accession IDs; data returned from UniProt REST API (scripts/uniprot_client.py).
Boundary markers: The skill does not implement specific delimiters for the LLM to ignore instructions inside the retrieved protein data.
Capability inventory: Network access via requests library for GET and POST operations.
Sanitization: The requests library performs standard URL encoding for parameters passed to the API.

uniprot-database