Nexus MCP — AI Security Tools

Most tools have distinct purposes: check_injection detects prompt injection, pii_scan masks personal data, sanitize_content cleans external content, and validate_rag_input provides a pass/fail gate. However, check_injection and validate_rag_input overlap slightly as both handle injection detection, though their descriptions clarify that one is for detailed analysis and the other for quick decisions. The two key-getting tools (get_trial_key and get_pii_guard_key) are clearly separate but could be confused due to similar naming and function.

The naming is mixed: check_injection, sanitize_content, and validate_rag_input follow a verb_noun pattern, but get_trial_key and get_pii_guard_key use a get_noun pattern, and pii_scan uses a noun_verb style. This inconsistency makes the set less predictable, though all names are readable and descriptive. The deviation from a single convention reduces clarity but doesn't render the tools chaotic.

With 6 tools, the count is well-scoped for an AI security server. Each tool serves a specific function in the domain, such as injection detection, PII scanning, content sanitization, and API key management. This number allows for comprehensive coverage without being overwhelming, fitting typical server sizes of 3-15 tools and ensuring each tool earns its place.

The tool set covers key AI security workflows: injection detection (check_injection, validate_rag_input), PII handling (pii_scan), content sanitization (sanitize_content), and API access (get_trial_key, get_pii_guard_key). Minor gaps exist, such as no explicit tool for logging or monitoring security events, but agents can work around this. The domain is well-covered for Japanese-language applications, with no dead ends in core operations.