facebook-ads-library-mcp-server
Fail
Audited by Gen Agent Trust Hub on Oct 8, 2026
Risk Level: HIGHEXTERNAL_DOWNLOADSREMOTE_CODE_EXECUTIONCOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [EXTERNAL_DOWNLOADS]: The skill requires users to clone a repository from an unverified GitHub organization (
proxy-intell/facebook-ads-library-mcp). - [REMOTE_CODE_EXECUTION]: Installation instructions command the immediate execution of
./install.shorinstall.batafter cloning the external repository, which is a pattern for running potentially malicious scripts without prior review. - [COMMAND_EXECUTION]: The skill documentation includes commands for installing dependencies (
pip install -r requirements.txt) and configuring environment files. - [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted content from the Facebook Ads Library (ad text, images, and videos) and passes it to AI models for analysis. This creates an attack surface where malicious advertising content could influence the agent's behavior.
- Ingestion points: External image URLs, video URLs, and ad text retrieved via the
get_meta_adsand analysis tools. - Boundary markers: The instructions do not define delimiters or warnings to prevent the agent from following instructions embedded within the analyzed media.
- Capability inventory: The MCP server has network access to Meta, ScrapeCreators, and Google APIs, and performs local file system operations for media caching.
- Sanitization: There is no evidence of filtering or sanitizing external content before it is processed by the AI models.
Recommendations
- AI detected serious security threats
Audit Metadata