The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill processes external clinical data to produce structured treatment plans, creating a surface for indirect prompt injection.
Ingestion points: Patient demographics, clinical diagnoses, and natural language descriptions for schematics as outlined in the Visual Enhancement and Core Capabilities sections of SKILL.md.
Boundary markers: The instructions do not specify the use of delimiters or boundary markers to isolate processed clinical data from the agent's core instructions.
Capability inventory: The skill employs the Bash tool to execute LaTeX compilers (pdflatex, xelatex) and several local Python scripts (scripts/generate_schematic.py, timeline_generator.py, check_completeness.py, validate_treatment_plan.py) as described in SKILL.md.
Sanitization: There is no mention of input sanitization or validation for the text interpolated into document templates or passed as arguments to the Python utility scripts.
[DYNAMIC_EXECUTION]: The skill generates and compiles executable LaTeX content and utilizes local scripts to process data at runtime.
Evidence: The skill instructions involve creating .tex files from templates and executing pdflatex or xelatex to generate PDF reports. It also describes a workflow for generating AI-powered schematics and timelines using Python scripts that take natural language input.
[COMMAND_EXECUTION]: The skill relies on the Bash tool to perform its primary functions.
Evidence: SKILL.md provides multiple shell command examples for compiling documents (pdflatex, xelatex, bibtex) and executing utility scripts (python scripts/generate_schematic.py, python check_completeness.py, etc.).

treatment-plans