The Agent Skills Directory

[SAFE]: The skill is designed for privacy compliance and implements several security best practices. The _redact_connection_string function in scripts/agent.py ensures that database credentials are not leaked into logs. Furthermore, the consolidation logic uses SHA-256 hashing to mask sensitive PII values while maintaining an audit trail.
[INDIRECT_PROMPT_INJECTION]: The skill's primary function is to ingest and process potentially untrusted data from various sources such as application logs, database records, and file exports.
Ingestion points: The PIIDiscoveryEngine in scripts/agent.py reads data from files and directories provided as input parameters.
Boundary markers: The PII discovery process does not utilize explicit boundary markers or delimiters when processing ingested text.
Capability inventory: The skill is limited to local file system reads and writes for logging and report generation. No network exfiltration or remote command execution capabilities were identified.
Sanitization: The skill includes an ExemptionReviewer and redaction logic to filter sensitive information before final response generation.

implementing-gdpr-data-subject-access-request