vaccine-schedule-tracker

Pass

Audited by Gen Agent Trust Hub on Sep 21, 2026

Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
  • [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted patient data, creating a potential surface for indirect prompt injection.
  • Ingestion points: The run_vaccine_tracker function in script.py and skill.py ingests user-supplied dob (date of birth), doses_received (vaccination history), and conditions (patient medical history).
  • Boundary markers: The skill's instructions do not define clear delimiters or "ignore instructions" wrappers for the patient-specific data provided at runtime.
  • Capability inventory: The skill is limited to deterministic Python logic for date calculation and keyword matching; it contains no subprocess, network, or file-writing capabilities across script.py or skill.py.
  • Sanitization: No sanitization is performed on input strings beyond basic case-insensitive keyword matching for contraindication screening.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 21, 2026, 11:46 AM