The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill processes large volumes of untrusted data from mobile device backups and extractions (e.g., SMS, messaging app databases, browser history), creating a significant attack surface for indirect prompt injection.
Ingestion points: iOSParser, AndroidParser, MessagingAnalyzer, and AppAnalyzer ingest data from provided evidence directories (e.g., /evidence/ios_backup/).
Boundary markers: The instructions do not specify any boundary markers or instructions for the agent to ignore embedded malicious content within the forensic artifacts.
Capability inventory: The skill can parse databases, extract files, and generate HTML reports based on the ingested content.
Sanitization: There is no evidence of sanitization or filtering for the external content before it is processed or included in reports.
[DATA_EXFILTRATION]: The skill is designed to access and extract highly sensitive data from mobile devices, which represents a high data exposure risk if misused.
Evidence: Capabilities include extract_cloud_tokens(), get_messages(), get_location_history(), and get_call_history(). These methods provide programmatic access to authentication tokens and extensive PII.

mobile-forensics