The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill provides logic for processing user data requests, creating a surface for potential injection via user-supplied identifiers.
Ingestion points: The user_id parameter in the ConsentManager class and the user_identifier parameter in the DSARProcessor class in SKILL.md.
Boundary markers: Absent. The provided logic does not include delimiters or specific instructions to the agent to treat these identifiers as untrusted data.
Capability inventory: The scripts include capabilities for database insertion (self.db.consent_records.insert), data extraction (source.extract_user_data), and data deletion (source.delete_user_data).
Sanitization: The implementation relies on standard hashing and JSON serialization but lacks explicit validation or escaping for the input identifiers.

gdpr-compliance