drug-interaction-checker
Pass
Audited by Gen Agent Trust Hub on Sep 23, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [SAFE]: The skill performs drug interaction checks using a purely local workflow. It references internal JSON files and hardcoded data structures within
scripts/main.pywithout initiating any network connections, external API calls, or remote data fetches. - [SAFE]: Code execution is limited to a local Python script that performs string matching and logical comparisons. The implementation does not utilize
subprocess,os.system, or any other functions capable of executing shell commands or interacting with the host system beyond basic standard output. - [INDIRECT_PROMPT_INJECTION]: The skill has an ingestion surface for untrusted data via the
--drugscommand-line argument inscripts/main.py. The risk is mitigated as the input is sanitized through normalization (lowercasing and character replacement) and used exclusively for dictionary key lookups. There is no interpolation of this data into executable code or unformatted prompts, and the output includes clear clinical disclaimers to prevent misuse of findings.
Audit Metadata