The Agent Skills Directory

[SAFE]: No malicious behavior, obfuscation, or safety bypass attempts were found. The skill documentation promotes best practices for pandas development, such as vectorization and memory optimization.
[REMOTE_CODE_EXECUTION]: Evaluation of dynamic execution surface. The reference material references/performance-optimization.md illustrates the use of pd.eval() and df.query() for efficient expression evaluation. While these features utilize dynamic execution, they are standard components of the pandas library used here for performance optimization.
[DATA_EXFILTRATION]: Indirect prompt injection attack surface evaluation.
Ingestion points: Data enters the agent's context through file reading operations like pd.read_csv(), as mentioned in references/data-cleaning.md and references/performance-optimization.md.
Boundary markers: No specific boundary markers or instructions to isolate untrusted data are present in the provided snippets.
Capability inventory: The skill provides capabilities for general data processing and analysis within a Python environment.
Sanitization: The skill suggests schema validation using the pandera library, though it does not explicitly address the prevention of malicious instructions embedded in data files.

pandas-pro