The Agent Skills Directory

[SAFE]: The skill uses deterministic Python code for all analytical tasks, relying on established libraries like pandas and scipy for statistical calculations and data manipulation.
[SAFE]: The prepare_for_exploration.py script implements a secure 'allow-list' approach for dynamic column operations. By mapping specific keywords to lambda functions, it explicitly avoids the use of eval() or exec(), preventing potential code injection vulnerabilities.
[SAFE]: Database exploration patterns correctly utilize environment variables (DATABASE_URL) for connection strings. This ensures that sensitive credentials are not hardcoded within the skill or instructions, adhering to secret management best practices.
[SAFE]: The skill enforces a 'read-only' constraint during the exploration phase, ensuring that raw data is not modified or deleted while performing analysis.
[SAFE]: Text processing logic in detect_patterns.py uses regular expressions to identify patterns (URLs, emails, phones) but does not execute or navigate to these strings, mitigating risks associated with malicious data content.

magic-data-exploration