The Agent Skills Directory

[SAFE]: No malicious patterns were detected. The skill demonstrates high security maturity through several proactive measures:
PHI Protection: Implements mandatory data privacy checks and warnings for Protected Health Information (PHI) before reading any tabular data files.
Code Quality Gate: Employs a specialized linter ('check_generated_code.py') to scan AI-emitted analysis code for non-reproducible patterns, hardcoded absolute paths (which pose a PII risk), and data integrity violations.
Path Sanitization: The execution runner ('run_analysis.py') implements strict validation of project-relative paths, effectively preventing directory traversal and symlink attacks.
Reproducibility Enforcement: Mandatory random seeding and environment metadata recording are required for all analysis scripts to ensure scientific integrity.
[INDIRECT_PROMPT_INJECTION]: The skill processes external research datasets (CSV, Excel) and uses this data to inform code generation, presenting a potential surface for indirect prompt injection.
Ingestion points: Phase 1 (Data Assessment) reads tabular data files.
Boundary markers: The skill maps data to statistical variables but lacks explicit delimiters for data content within the LLM context.
Capability inventory: The skill utilizes the 'Bash' tool for script execution and the 'Write' tool to save artifacts.
Sanitization: The skill implements a robust mitigation via 'check_generated_code.py', which validates the emitted code for security and reproducibility issues before it is executed.
[DYNAMIC_EXECUTION]: The skill generates and executes Python and R scripts to perform complex statistical tasks. This capability is safely managed through the use of bundled templates and a mandatory automated linter that scans for insecure code patterns and ensures all randomness is correctly seeded. Dynamic imports in utility scripts are restricted to hardcoded package lists or internal skill paths for version auditing purposes.

analyze-stats