The Agent Skills Directory

[COMMAND_EXECUTION]: The skill documentation instructs the agent to execute local Python scripts located in a scripts/ directory, such as scripts/risk_assessment.py and scripts/compliance_checker.py. This involves the use of the shell to perform information security management tasks.
[INDIRECT_PROMPT_INJECTION]: The skill processes untrusted external data which could potentially contain malicious instructions intended to influence the agent's output or behavior.
Ingestion points: Untrusted data enters the agent context through external files like assets.csv and controls.csv, which are passed as arguments to the execution scripts.
Boundary markers: The skill instructions do not specify any delimiters or safety prompts to prevent the agent from obeying instructions embedded within the data files.
Capability inventory: The skill involves executing Python scripts that perform file system read and write operations, such as generating risk_register.json or gap_analysis.md reports.
Sanitization: There is no evidence of data sanitization or validation performed on the input files before they are processed by the scripts.

information-security-manager-iso27001