tools-mcp-bridges
Pass
Audited by Gen Agent Trust Hub on Oct 15, 2026
Risk Level: SAFEEXTERNAL_DOWNLOADSCOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTIONDYNAMIC_EXECUTION
Full Analysis
- [COMMAND_EXECUTION]: The skill is designed to enable arbitrary shell command execution by the AI agent through the 'mcp-kali-server' terminal bridge. The documentation explicitly acknowledges this as an arbitrary command execution surface.
- [EXTERNAL_DOWNLOADS]: The skill recommends installing and running software from community repositories, including a Docker image (
ramkansal/pentestmcp), various Python packages via pip (mcp-security-hub), and Node.js packages via npm. - [DYNAMIC_EXECUTION]: The generic terminal bridge functionality involves executing dynamically generated strings as shell commands on the host system or within a container environment.
- [INDIRECT_PROMPT_INJECTION]: The skill describes a workflow that ingests outputs from various security tools (e.g., nmap, nuclei, ZAP). These outputs could potentially contain adversarial content designed to influence the agent's behavior.
- Ingestion points: Tool outputs retrieved through MCP servers (SKILL.md).
- Boundary markers: None explicitly defined in the provided configuration.
- Capability inventory: Full terminal command execution, network scanning, and exploitation tool usage (nmap, Metasploit, etc.).
- Sanitization: No evidence of input validation or output sanitization before the data is processed by the agent.
Audit Metadata