The Agent Skills Directory

[SAFE]: Extensive analysis of the skill's instructions, metadata, and accompanying scripts revealed no evidence of prompt injection, obfuscation, or malicious intent. The skill adheres to its stated purpose of professional medical documentation.
[EXTERNAL_DOWNLOADS]: The skill relies on standard, reputable Python packages (pandas, numpy, scipy, lifelines, matplotlib, pyyaml) from official registries. These libraries are standard requirements for scientific and statistical computing.
[COMMAND_EXECUTION]: The skill uses the Bash tool for the legitimate purpose of executing its local Python analysis scripts. The scripts are transparent and perform expected statistical operations on provided clinical datasets.
[INDIRECT_PROMPT_INJECTION]: The skill possesses a data ingestion surface through the processing of CSV and JSON clinical records.
Ingestion points: Scripts in the scripts/ directory load external data files provided by the user.
Boundary markers: Data is processed via standard CSV/JSON loaders and interpolated into templates; no specific LLM boundary markers are implemented within the data ingestion scripts.
Capability inventory: The skill uses Read, Write, Edit, and Bash tools to analyze data and produce output files.
Sanitization: The provided scripts/validate_cds_document.py tool performs post-generation validation to ensure HIPAA compliance and citation integrity.

clinical-decision-support