superseller-content-factory
Pass
Audited by Gen Agent Trust Hub on May 18, 2026
Risk Level: SAFEDATA_EXFILTRATIONEXTERNAL_DOWNLOADSCOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
- [PROMPT_INJECTION]: Indirect prompt injection surface detected in the content research stage.
- Ingestion points: The skill pulls competitor highlights and trending themes from external sources (Apify scrapes) in Stage 1.
- Boundary markers: No specific delimiters or instructions to ignore embedded commands within the scraped data are present in the logic.
- Capability inventory: The skill has the ability to write to the ig_content_calendar database and execute video generation via MCP tools.
- Sanitization: There is no documented validation or sanitization of the external data before it is synthesized into content briefs.- [DATA_EXFILTRATION]: The skill accesses a PostgreSQL database to retrieve tenant-specific strategy, brand metadata, and actor information. It also transmits content previews and metadata to external WhatsApp approval groups.- [EXTERNAL_DOWNLOADS]: The skill initiates external data retrieval by pulling competitor data and metrics via third-party scraping services.- [COMMAND_EXECUTION]: The skill executes actions through MCP tools for video generation and credit management, which involve external API interactions and financial expenditure.
Audit Metadata