The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill contains logic for parsing and processing external medical data (HL7 v2 messages and FHIR resources), which serves as a vulnerability surface for indirect prompt injection attacks if the ingested content contains adversarial instructions.
Ingestion points: The parse_hl7_message function in SKILL.md and the EHRSystem class in references/ELECTRONIC_HEALTH_RECORD_SYSTEM.md process externally supplied strings and medical records.
Boundary markers: There are no explicit boundary markers or system instructions to ignore potential commands embedded within the medical data fields (e.g., patient notes, names, or HL7 segments).
Capability inventory: The skill has access to file system operations through allowed-tools (Read, Write, Edit) and performs cryptographic operations using the cryptography library.
Sanitization: The provided Python examples demonstrate data parsing and encryption but lack explicit sanitization or validation to filter out non-medical instructional content from input strings.

healthcare-expert