The Agent Skills Directory

[COMMAND_EXECUTION]: The skill uses a Python script (scripts/profiler.py) to perform its main functionality of scanning datasets and generating reports. This is an expected behavior for a data profiling tool.
[EXTERNAL_DOWNLOADS]: The script depends on standard, widely-used Python packages including pandas and numpy. These are recognized as well-known libraries in the data science ecosystem.
[PROMPT_INJECTION]: A surface for indirect prompt injection exists because the skill processes untrusted user data and incorporates it into a report that may be read by an agent.
Ingestion points: Data is ingested via the --data_path parameter in scripts/profiler.py using pd.read_csv and pd.read_parquet.
Boundary markers: The markdown report generation does not implement specific delimiters or instructions to treat data-derived strings as non-executable content.
Capability inventory: The skill uses a ReportWriter class (imported from a system-level utility directory) to write files and emit state updates.
Sanitization: The script performs basic string conversion and truncation on sample data and column names before inclusion in the output report.

data-profiling