The Agent Skills Directory

[SAFE]: The skill provides comprehensive regulatory documentation and local diagnostic scripts intended to help medical device companies navigate FDA and HIPAA requirements. No malicious patterns, obfuscation, or unauthorized access attempts were detected.- [DATA_EXPOSURE_AND_EXFILTRATION]: The provided Python scripts (fda_submission_tracker.py, qsr_compliance_checker.py, and hipaa_risk_assessment.py) perform read-only operations on local project files specified by the user. There are no network-capable libraries (like requests or urllib) or exfiltration patterns present. The scripts contain regex patterns that search for hardcoded secrets as a security check, which is a safe diagnostic practice.- [INDIRECT_PROMPT_INJECTION]: The skill possesses a data ingestion surface as it analyzes external project files. * Ingestion points: scripts/hipaa_risk_assessment.py, scripts/qsr_compliance_checker.py, and scripts/fda_submission_tracker.py ingest data from a user-specified directory. * Boundary markers: Absent. * Capability inventory: Read-only file system access via pathlib.Path.glob and read_text. No shell execution, write operations, or network access. * Sanitization: None; the scripts perform direct regex matching on raw file content. The impact of processing untrusted data is negligible as the scripts do not execute the content or transmit it.- [COMMAND_EXECUTION]: No dangerous command execution patterns (such as subprocess.run with shell=True or os.system with user input) were found. The usage examples in SKILL.md demonstrate standard local execution of the provided Python scripts.

fda-consultant-specialist