The Agent Skills Directory

[PROMPT_INJECTION]: The skill ingests data from CSV files. While the processing is primarily numerical, the column names are included in the JSON output, creating a theoretical indirect prompt injection surface. Evidence: The summarize function in scripts/risk_measurement_diagnostics.py includes list(working_frame.columns) in the JSON payload.
[COMMAND_EXECUTION]: The skill executes a local Python script to perform data diagnostics, which is the core intended functionality of the skill.
[EXTERNAL_DOWNLOADS]: The skill uses the pandas library, a well-known and standard industry tool for data analysis.

risk-measurement