The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill utilizes a Python script to process external data which could potentially contain malicious instructions intended to influence the agent's behavior during audit planning.
Ingestion points: The scripts/audit_schedule_optimizer.py script reads and parses user-provided data from a JSON file specified via the --processes command-line argument.
Boundary markers: There are no explicit boundary markers or instructions within the script to disregard or sanitize natural language instructions embedded within the JSON fields.
Capability inventory: The script is restricted to performing scheduling calculations and printing the results to the console. It does not include network access, file system write capabilities, or the ability to execute system commands.
Sanitization: While the script validates the risk_level field against an internal enumeration, other fields such as name and iso_clause are treated as raw strings and included in the output report.

qms-audit-expert