The Agent Skills Directory

[SAFE]: The skill provides a robust framework for clinical documentation. It uses local LaTeX templates and Python scripts (check_completeness.py, validate_treatment_plan.py, timeline_generator.py, generate_template.py) to automate the creation and validation of medical plans. All scripts were analyzed and found to perform benign operations such as regex-based file parsing and local file copying.
[SAFE]: The skill contains a reference to the scientific-schematics skill and instructs the execution of scripts/generate_schematic.py. While this file is not included in the provided package, it is treated as a vendor-internal resource and does not pose a security risk. No external network downloads or unverifiable third-party code executions were identified.
[PROMPT_INJECTION]: The skill processes patient data provided by the user to populate LaTeX templates. This creates a surface for indirect prompt injection if malicious instructions are embedded in the patient data. However, the risk is reduced as the processing is limited to regex validation and compilation into PDF format.
Ingestion points: Patient demographic information and clinical history (e.g., ICD-10 codes, diagnoses, medication lists) entered by the user into the [bracketed placeholders] of the LaTeX templates.
Boundary markers: None identified in the provided templates or instructions.
Capability inventory: File system writes (Write tool) to create .tex and .pdf files, and command execution (Bash tool) to run Python scripts and the pdflatex compiler.
Sanitization: No explicit sanitization or escaping of user input was found in the provided Python scripts or instructions.

alterlab-treatment-plans