The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill processes external clinical data and identifiers which could potentially be used for indirect prompt injection attacks.\n
Ingestion points: The skill ingests data through pacs_query_retrieve.py (via command-line arguments), audit_osirix_study.py (by scanning DICOM file metadata in user-provided directories), and export_key_images.scpt (processing patient identifiers).\n
Boundary markers: No explicit delimiters or instructions are provided to distinguish between trusted commands and untrusted data within the agent's context.\n
Capability inventory: The skill facilitates network operations using pynetdicom (DICOM C-FIND and C-MOVE operations), local database management via sqlite3, and file system interactions such as JPEG exports.\n
Sanitization: There are no built-in validation or sanitization mechanisms to filter clinical metadata before it is used in automated workflows.\n- [EXTERNAL_DOWNLOADS]: The skill suggests the installation of well-known and reputable Python libraries specialized for medical imaging.\n
Python packages: pydicom and pynetdicom are required for the DICOM networking and auditing scripts.\n- [COMMAND_EXECUTION]: The skill provides instructions and scripts that execute various system commands for workstation automation and maintenance.\n
Database maintenance is performed using sqlite3 to optimize the OsiriX SQLite database (VACUUM, ANALYZE).\n
Workflow automation is achieved through osascript, which executes AppleScripts to interact with the OsiriX application.\n
System diagnostic tools like lsof, open, and curl are utilized to manage network ports and trigger URL schemes.

osirix