The Agent Skills Directory

[DYNAMIC_EXECUTION]: The skill generates and executes Python code at runtime to perform disproportionality calculations and statistical signal detection. It utilizes standard data science libraries including pandas, scipy, statsmodels, and matplotlib to process drug safety data.
[COMMAND_EXECUTION]: The agent uses shell commands to run the dynamically generated Python analysis scripts and to manage report file creation.
[INDIRECT_PROMPT_INJECTION]: The skill processes clinical and safety data from external sources, presenting an attack surface for indirect prompt injection.
Ingestion points: Data is ingested from the FDA Adverse Event Reporting System (FAERS), DailyMed (NIH), and PharmGKB clinical annotations via ToolUniverse tools.
Boundary markers: The report templates use markdown headers to structure data, but do not specify explicit delimiters or "ignore instructions" prompts for tool output.
Capability inventory: The skill has the ability to write local files (markdown, CSV) and execute Python code via a shell environment.
Sanitization: The instructions do not describe specific sanitization or filtering of external data content prior to its inclusion in reports or code execution contexts.
[EXTERNAL_DOWNLOADS]: The skill references and fetches data from established governmental and scientific organizations, including the FDA (open.fda.gov), the National Institutes of Health (dailymed.nlm.nih.gov, clinicaltrials.gov), and research databases like PharmGKB and KEGG. These resources are well-known and trusted in the medical research community.

tooluniverse-pharmacovigilance