The Agent Skills Directory

[PROMPT_INJECTION]: No instructions found that attempt to override agent behavior, bypass safety filters, or extract system prompts.
[DATA_EXFILTRATION]: No evidence of unauthorized data access, network exfiltration, or hardcoded credentials. The included scripts process local CSV and JSON files provided via command-line arguments and do not perform network operations.
[REMOTE_CODE_EXECUTION]: No remote script downloads or execution patterns (such as curl | bash) were detected. All provided scripts use only Python's standard library.
[OBFUSCATION]: No obfuscated content, encoded commands (Base64/Hex), or hidden characters were found in the skill files.
[COMMAND_EXECUTION]: The skill documents the use of local Python scripts for data processing. These scripts perform standard data aggregation and validation tasks without executing arbitrary shell commands or user-supplied input strings.
[INDIRECT_PROMPT_INJECTION]: The skill identifies a surface for indirect prompt injection as it processes external CSV and JSON data that is subsequently interpreted by the agent to generate insights. However, the logic is limited to standard data analysis and includes validation steps (metric_validator.py) which act as a security control for data integrity.

business-intelligence