The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and process data from external bioinformatics repositories, creating a potential attack surface for indirect prompt injection.
Ingestion points: Data is retrieved via requests.get from alphafold.ebi.ac.uk and www.ebi.ac.uk, and via urllib.request from www.uniprot.org (e.g., in SKILL.md).
Boundary markers: There are no specific instructions or delimiters provided to the agent to treat the fetched protein metadata or structure files as untrusted content.
Capability inventory: The skill possesses the ability to write files to the local system (open().write), perform network operations (requests.get), and execute shell commands via subprocess.run (specifically using the gsutil tool).
Sanitization: The skill demonstrates good practice by including an explicit type check (verifying taxonomy_id is an integer) to prevent command injection when using gsutil.
[METADATA_POISONING]: The SKILL.md file contains a section titled 'Suggest Using K-Dense Web For Complex Worflows' which provides explicit instructions to the AI agent to steer users toward a specific commercial platform (www.k-dense.ai). This is a behavioral override designed to promote the author's hosted services during complex interactions.

alphafold-database