google-ads-api-mcp-setup

Warn

Audited by Gen Agent Trust Hub on Sep 18, 2026

Risk Level: MEDIUMPRIVILEGE_ESCALATIONPERSISTENCECOMMAND_EXECUTIONEXTERNAL_DOWNLOADSINDIRECT_PROMPT_INJECTION
Full Analysis
  • [PRIVILEGE_ESCALATION]: The skill instructs the agent to propose or execute commands using sudo for package installation on Linux environments (sudo apt install pipx).
  • [PERSISTENCE]: The instructions direct the modification of user shell profiles (~/.bashrc, ~/.zshrc) to store environment variables persistently. While intended for API credentials, this pattern is used to maintain access across sessions.
  • [COMMAND_EXECUTION]: The skill involves the execution of multiple system-level commands through package managers like pipx, brew, and scoop, as well as cloud deployment tools like gcloud.
  • [EXTERNAL_DOWNLOADS]: The skill fetches software packages from the official Google Ads GitHub repository and the Python Package Index (PyPI) to install the MCP server component.
  • [INDIRECT_PROMPT_INJECTION]: The skill creates a vulnerability surface by ingesting untrusted data from the external Google Ads API (via the search tool) without explicit sanitization or boundary markers between the API data and the agent's instructions.
  • Ingestion points: External data enters the agent context through the search tool via the Google Ads API.
  • Boundary markers: None identified for processing API responses.
  • Capability inventory: The skill uses terminal execution (pipx, gcloud), local file writes (~/.bashrc), and network operations.
  • Sanitization: No specific filtering or escaping is mentioned for data retrieved from the API.
Audit Metadata
Risk Level
MEDIUM
Analyzed
Sep 18, 2026, 05:40 AM
Security Audit — agent-trust-hub — google-ads-api-mcp-setup