The Agent Skills Directory

[REMOTE_CODE_EXECUTION]: The skill provides instructions in references/brenda.md for the agent to generate and execute a local Python script using the zeep library to interact with a SOAP API. This involves dynamic code generation and runtime execution within the agent environment.
[EXTERNAL_DOWNLOADS]: The skill facilitates data retrieval from 78 distinct external domains, including government agencies (NASA, NIST, USGS, EPA, FDA, USPTO, SEC), international organizations (WHO, World Bank, Eurostat, ECB), and major scientific research repositories (EBI, NCBI, UniProt, RCSB).
[PROMPT_INJECTION]:
Ingestion points: Raw data (JSON, XML, and flat-files) is ingested from 78 external API endpoints (SKILL.md).
Boundary markers: Not present. The agent is instructed to return raw JSON results without explicit validation or delimiters between data and instructions.
Capability inventory: Access to platform-specific fetch tools (WebFetch, web_fetch, read_url_content), curl via shell, and a Python interpreter for script generation (references/brenda.md).
Sanitization: Absent. The skill assumes the reliability of data provided by the external database vendors.

database-lookup