medication-administration-safety
Pass
Audited by Gen Agent Trust Hub on Sep 21, 2026
Risk Level: SAFENO_CODEINDIRECT_PROMPT_INJECTIONEXTERNAL_DOWNLOADS
Full Analysis
- [NO_CODE]: The skill consists exclusively of markdown documentation and logic instructions for the agent. It does not include any Python scripts, JavaScript files, or binary executables.
- [INDIRECT_PROMPT_INJECTION]: The skill is designed to process untrusted data such as patient identifiers and medication orders, posing a potential surface for indirect instructions.
- Ingestion points: Clinical data including patient MRNs, drug names, and dosages are processed through user prompts.
- Boundary markers: The skill does not define boundary markers or delimiters to separate instructions from the data being analyzed.
- Capability inventory: The skill is restricted to generating text-based safety summaries and checklists; it possesses no file system or network access capabilities.
- Sanitization: No input validation or filtering mechanisms are specified for the ingested clinical data.
- [EXTERNAL_DOWNLOADS]: The quick install guide references the vendor's repository (Open-Medica/open-medical-skills) via an installation command.
Audit Metadata