Zero-dependency web security auditing, strict CSP generator, SRI hasher, WCAG contrast analyzer, and MCP server
Server has 7 tools with complete input schemas and descriptions. Naming follows verb_noun convention (audit_security, calculate_contrast, generate_*). Descriptions are substantive (100-300 chars), exceeding the 10-char minimum. However, output schemas are not documented in the visible code, responses are described in text but not formally specified. Error handling is minimal; no recovery guidance or actionable error messages visible. Parameter descriptions are present but lack constraint details (ranges, patterns, validation rules). One tool (inject_sri_into_html) is marked WRITE but lacks confirmation/dry-run pattern. Overall: solid foundation, but incomplete schema documentation and weak error handling prevent higher scores.
Audits web security headers, cookies, CORS configuration, and dark patterns. Supports live URLs, raw headers, HTML files, and web server configurations (Netlify _headers, Vercel vercel.json, Nginx nginx.conf, Caddy Caddyfile, Apache .htaccess). Returns security score (0-100), grade (A+ to F), and detailed findings with remediation hints.
Calculates WCAG 2.2 color contrast ratio between foreground and background colors. Supports hex (#RGB, #RRGGBB, #RRGGBBAA), rgb/rgba(), and named colors. Returns contrast ratio, compliance status (AA/AAA), and accessibility suggestions.
Generates a strict Content Security Policy (CSP Level 3) with nonce support, 'strict-dynamic', and defense-in-depth directives. Supports custom script/style sources, external domains, and development mode (unsafe-eval for HMR).
Generates MCP client configuration files for Claude Desktop, Cursor, Cline, Zed, and generic MCP hosts. Outputs JSON/YAML configs with stdio transport and tool definitions.
Generates hardened security configuration files for multiple platforms: Netlify (_headers), Vercel (vercel.json), Nginx (nginx.conf), Caddy (Caddyfile), Apache (.htaccess), and Next.js (middleware.ts). Includes CSP, HSTS, CORS, and cross-origin isolation headers.
Output schemas not formally documented. Tool descriptions state what is returned (e.g., 'Returns security score, grade, findings') but no JSON Schema or structured type definitions are visible in code. LLMs cannot plan downstream tool calls without knowing response field names and types.
Error handling lacks recovery guidance. No visible error classification (retryable vs. user-fixable vs. fatal) or actionable error messages. E.g., if audit_security fails on a malformed URL, the LLM receives no hint on what to try next.
Parameter constraints not fully specified. E.g., font_size_pt in calculate_contrast lacks min/max bounds; platforms array in generate_remediation_configs lacks explanation of 'all' behavior; no regex patterns or length limits documented in descriptions.
| Scored | Grade | Overall | Spec posture | Rubric |
|---|---|---|---|---|
| 2026-09-23 | C | 61 | 2026-07-28+ | v2 |
Computes Subresource Integrity (SRI) cryptographic hash for scripts and stylesheets. Supports SHA-256, SHA-384, and SHA-512 algorithms. Returns integrity string, HTML tag templates, and hex/base64 digests.
Rewrites HTML <script> and <link> tags to include Subresource Integrity (SRI) hashes. Fetches remote resources, computes SHA-384 hashes, and injects integrity attributes with crossorigin='anonymous'.
inject_sri_into_html marked WRITE but no confirmation/dry-run pattern. Modifying HTML is destructive; agents should preview changes or confirm before execution to prevent accidental data loss.
Pagination and result limits not documented. audit_security and generate_remediation_configs may return large result sets; no mention of limits, pagination, or how to handle truncation.