The Agent Skills Directory

Prompt Injection (SAFE): No instructions attempting to override agent behavior or bypass safety filters were found. The content is strictly focused on medical documentation standards.
Data Exfiltration & Credentials (SAFE): No hardcoded secrets, API keys, or credentials detected. The scripts operate exclusively on local files provided by the user and do not perform any network operations (no use of requests, urllib, or socket).
Remote Code Execution & Dependencies (SAFE): The provided scripts (compliance_checker.py, extract_clinical_data.py, etc.) use only Python standard libraries (argparse, json, re, shutil, csv, pathlib). There are no external dependencies or calls to eval(), exec(), or subprocess.
Obfuscation (SAFE): No Base64 encoding, zero-width characters, or other obfuscation techniques were detected. The code and markdown files are clear and readable.
Indirect Prompt Injection (LOW/SAFE): While the skill is designed to process external clinical data (reports and CSVs), the scripts are limited to regex-based extraction and formatting. There is no execution path where data processed from these files could influence the agent's logic or lead to dangerous operations.

clinical-reports