The Agent Skills Directory

[DATA_EXFILTRATION]: The skill contains Python code patterns that utilize the requests library to fetch data from user-defined or dynamic URLs (e.g., requests.get(url) and requests.get(api_url)). While these are intended for legitimate data acquisition, they represent a capability to perform network operations to non-whitelisted domains.
[INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and process data from external sources (CSV, JSON, TSV, SAS) which could potentially contain malicious content intended to influence the agent's output during the generation of the final statistical report.
Ingestion points: The download_and_parse helper function and the REST API pagination block in SKILL.md are primary points where untrusted data enters the agent context.
Boundary markers: The instructions do not define specific delimiters or instructions for the agent to ignore potentially malicious embedded content within the datasets.
Capability inventory: The skill has the capability to perform network operations via requests and interact with the file system through standard Python libraries.
Sanitization: No explicit data sanitization or validation logic is provided for the downloaded external content.
[DYNAMIC_EXECUTION]: The skill facilitates the generation and execution of Python scripts based on templates provided for statistical analysis. While these templates are for standard epidemiological methods, the runtime assembly of code for regression and sensitivity analysis constitutes dynamic execution based on data and user inputs.

tooluniverse-epidemiological-analysis