The Agent Skills Directory

[PROMPT_INJECTION]: The skill processes external datasets, creating a surface for indirect prompt injection where malicious instructions within data files could influence agent behavior.
Ingestion points: Data is loaded from files like data.csv and unity_profiler_output.csv using pd.read_csv.
Boundary markers: The instructions lack markers or delimiters to differentiate data from instructions.
Capability inventory: Tools enabled include Bash, Read, Grep, and Glob, and the code can perform file system operations.
Sanitization: No input validation or sanitization is present for the processed files.

data-analysis