data-export

Warn

Audited by Gen Agent Trust Hub on Apr 1, 2026

Risk Level: MEDIUMDATA_EXFILTRATION
Full Analysis
  • [DATA_EXFILTRATION]: The skill is designed to read sensitive marketing data, brand profiles, and campaign information from the local file system (specifically within the ~/.claude-marketing/ directory) and transmit this information to external endpoints.- [DATA_EXFILTRATION]: It facilitates the movement of data to user-specified destinations including BigQuery, Google Sheets, and Supabase, which creates a potential vector for data exfiltration if destinations are controlled by a malicious actor.- [DATA_EXFILTRATION]: The skill handles access permissions and credentials for external database services (Supabase) to perform data uploads.- [DATA_EXFILTRATION]: (Indirect Prompt Injection Surface):\n
  • Ingestion points: Reads marketing data from local JSON files (campaign-tracker.json, execution-tracker.json) and connected MCPs (Google Analytics, CRM, ad platforms).\n
  • Boundary markers: No explicit XML/markdown delimiters or 'ignore embedded instructions' warnings are defined for the data being read and transformed.\n
  • Capability inventory: Performs network exports to cloud services and writes execution logs to the local file system.\n
  • Sanitization: Includes PII redaction and data validation logic, but lacks specific mechanisms to prevent the execution of instructions that may be embedded within the ingested marketing data.
Audit Metadata
Risk Level
MEDIUM
Analyzed
Apr 1, 2026, 01:18 AM