The Agent Skills Directory

[SAFE]: The skill facilitates the installation and usage of the PyTDC library, which is a standard, well-known scientific package for pharmaceutical and biochemical machine learning.
[INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and process large-scale datasets from external sources (Therapeutics Data Commons). While this is the intended function, it creates a potential surface for indirect prompt injection if the agent processes untrusted data without sanitization.
Ingestion points: Data loading methods in tdc.single_pred, tdc.multi_pred, and tdc.generation as demonstrated in SKILL.md and template scripts.
Boundary markers: The instructions do not specify explicit delimiters or markers to differentiate dataset content from agent instructions.
Capability inventory: Network access for dataset downloads and file system access for storing data in the data/ directory.
Sanitization: No specific content validation or sanitization of the downloaded chemical or therapeutic data is implemented in the provided templates.

pytdc