Security Operations MCP Server — turns AI agents into security analysts. Scan HTTP headers, analyze SSL/TLS, enumerate DNS, fingerprint tech stacks, and generate full security reports.
Sentinel MCP has well-structured tool definitions with clear descriptions and proper Zod schemas. All 6 tools are explicitly registered with descriptions (194 - 280 chars, within baseline 34 - 392 range) and input schemas. However, output schemas are not documented, responses are JSON-stringified without declared structure. Parameter descriptions are present but lack constraint details (enums, ranges, formats). Error handling returns generic messages without recovery guidance. No tool annotations (readOnlyHint, destructiveHint) despite all tools being read-only security scans. Tool names follow verb_noun pattern well (http_security_scan, ssl_certificate_check, dns_recon, port_scan, tech_fingerprint, security_report).
Perform DNS reconnaissance on a domain. Enumerates A, AAAA, MX, TXT, NS, CNAME, and SOA records. Checks email security (SPF, DMARC, DKIM), nameserver redundancy, and CAA records. Identifies email spoofing vulnerabilities.
Analyze HTTP security headers of a target URL. Checks for HSTS, CSP, X-Frame-Options, and 10+ security headers. Returns a security grade (A+ to F), detailed findings, and remediation recommendations.
Scan common ports on a target host to identify open services and assess attack surface. Checks 30 well-known ports including databases (MySQL, PostgreSQL, MongoDB, Redis), remote access (SSH, RDP, Telnet), web servers, and more. Maps findings to risk levels and provides security recommendations. Only use for authorized targets.
Generate a comprehensive security assessment report for a target. Runs ALL scans (HTTP headers, SSL/TLS, DNS, ports, technology) concurrently and produces an executive summary with overall grade, critical findings, MITRE ATT&CK technique mapping, and prioritized remediation recommendations. This is the primary tool for a complete security audit.
Analyze SSL/TLS configuration of a target. Checks certificate validity, expiration, trust chain, protocol version (TLS 1.2/1.3), and cipher strength. Returns security grade and detailed findings.
Output schemas not documented. Tools return JSON-stringified results without declared field structure, forcing LLMs to parse unstructured text and guess at response shape for downstream chaining.
Parameter constraints not specified. 'url', 'target', 'domain' parameters lack format validation (URL vs hostname vs IP), length limits, or examples of valid input. port_scan 'timeout' has no min/max bounds.
Error messages lack recovery guidance. All tools return generic 'Error scanning X: <message>' without suggesting next steps (e.g., 'Invalid URL format. Try https://example.com or example.com').
| Scored | Grade | Overall | Spec posture | Rubric |
|---|---|---|---|---|
| 2026-09-23 | D | 59 | 2026-07-28+ | v2 |
Fingerprint the technology stack of a website. Detects web servers, frameworks, CMS platforms, JavaScript libraries, CDNs, WAFs, analytics tools, and more from HTTP headers and HTML content. Identifies security implications of detected technologies.
Tool annotations missing. All tools are read-only security scans but lack readOnlyHint annotation in schema. This prevents clients from optimizing caching, retry logic, or UI presentation.
port_scan optional parameters lack defaults. 'ports' and 'timeout' are optional but no default behavior is documented. Does omitting 'ports' scan 30 common ports? What is the default timeout?