mcp-security-hub
Warn
Audited by Gen Agent Trust Hub on Sep 18, 2026
Risk Level: MEDIUMEXTERNAL_DOWNLOADSPRIVILEGE_ESCALATIONCOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [EXTERNAL_DOWNLOADS]: The skill directs users to clone and build a software suite from a third-party GitHub repository (github.com/FuzzingLabs/mcp-security-hub) which is not verified as a trusted vendor.\n- [PRIVILEGE_ESCALATION]: The documentation suggests mounting the sensitive host Docker socket (/var/run/docker.sock) into the Trivy container to facilitate container scanning. Access to the Docker socket allows for interaction with the Docker daemon, which can be leveraged to gain information about or control over the host system.\n- [COMMAND_EXECUTION]: The Nmap MCP server includes a 'scan_custom' tool that permits the execution of arbitrary Nmap commands, granting the agent broad and flexible control over network-level operations.\n- [INDIRECT_PROMPT_INJECTION]: The skill integrates numerous tools designed to ingest and process untrusted data from external targets, such as network responses, web page content, and binary files, without implementing explicit boundary markers.\n
- Ingestion points: Data is ingested from network targets (Nmap, Nuclei), binary files (Radare2), and external repositories (Gitleaks) as described in SKILL.md.\n
- Boundary markers: The instructions do not define delimiters or specific 'ignore embedded instructions' warnings for the agent when processing data returned by the security tools.\n
- Capability inventory: The skill provides access to over 300 tools across 38 categories, including network exploit, database dumping, and binary analysis capabilities (SKILL.md).\n
- Sanitization: There is no evidence of output sanitization or validation of the content retrieved from external sources before it is presented to the agent.
Audit Metadata