The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill ingests untrusted input strings such as processing activity names, descriptions, and data elements which are then used to calculate risk scores and generate remediation plans. This represents a potential surface for indirect injection if the resulting reports are consumed by other automated agents.
Ingestion points: The register_processing_activity and map_data_flows functions in scripts/agent.py accept arbitrary user strings.
Boundary markers: None are implemented to delimit user-provided data from engine instructions.
Capability inventory: The skill has file-writing capabilities through the generate_dpia_report function in scripts/agent.py.
Sanitization: No sanitization or escaping is performed on user-provided strings before they are processed or written to reports.
[COMMAND_EXECUTION]: The skill provides the ability to write generated assessment reports to a user-specified file path. While this is an intended feature for document generation, it involves writing to the local file system based on an unvalidated path argument.
Evidence: The generate_dpia_report function in scripts/agent.py uses open(output_path, "w") to save assessment results.

performing-privacy-impact-assessment