DollhouseMCP

/** * Security Rules - Defines security patterns and checks * Based on OWASP Top 10, CWE Top 25, and DollhouseMCP-specific security requirements */ import type { SecurityRule, SecurityFinding } from '../types.js'; export class SecurityRules { /** * OWASP Top 10 security rules */ getOWASPRules(): SecurityRule[] { return [ { id: 'OWASP-A01-001', name: 'Hardcoded Secrets', description: 'Potential hardcoded secret or API key detected', severity: 'critical', category: 'code', pattern: /(?:api[_-]?key|secret|password|token|private[_-]?key)\s*[:=]\s*["'][a-zA-Z0-9+/=_-]{10,}["']/gi, remediation: 'Use environment variables or secure key management services instead of hardcoding secrets', references: ['https://owasp.org/Top10/A01_2021-Broken_Access_Control/'], tags: ['high-confidence'] }, { id: 'OWASP-A03-001', name: 'SQL Injection', description: 'Potential SQL injection vulnerability', severity: 'critical', category: 'code', pattern: /(?:query|execute)\s*\(\s*['"`].*\$\{[^}]+\}.*['"`]|['"`].*\+\s*[a-zA-Z_]\w*\s*\+.*['"`]\s*\)/g, remediation: 'Use parameterized queries or prepared statements', references: ['https://owasp.org/Top10/A03_2021-Injection/'] }, { id: 'OWASP-A03-002', name: 'Command Injection', description: 'Potential command injection vulnerability', severity: 'critical', category: 'code', pattern: /(?:exec|spawn|execSync|spawnSync)\s*\([^)]*(?:\$\{[^}]+\}|\+\s*[a-zA-Z_]\w*)/g, remediation: 'Validate and sanitize all user input before using in system commands', references: ['https://owasp.org/Top10/A03_2021-Injection/'] }, { id: 'OWASP-A03-003', name: 'Path Traversal', description: 'Potential path traversal vulnerability', severity: 'high', category: 'code', pattern: /(?:readFile|writeFile|readdir|mkdir|rm|unlink)[^(]*\([^)]*(?:\.\.[/\\].*\+|\+.*\.\.[/\\])/g, remediation: 'Validate and sanitize file paths, use path.resolve() and check against allowed directories', references: ['https://owasp.org/Top10/A03_2021-Injection/'] }, { id: 'OWASP-A03-004', name: 'XSS - Direct HTML Injection', description: 'Potential XSS vulnerability through direct HTML injection', severity: 'high', category: 'code', pattern: /innerHTML\s*=\s*[^'"`]*\$\{|dangerouslySetInnerHTML/g, remediation: 'Use textContent or proper HTML escaping functions', references: ['https://owasp.org/Top10/A03_2021-Injection/'] }, { id: 'OWASP-A05-001', name: 'Insecure Configuration', description: 'Security-sensitive configuration detected', severity: 'medium', category: 'code', pattern: /(?:NODE_TLS_REJECT_UNAUTHORIZED|strictSSL|rejectUnauthorized)\s*[:=]\s*(?:false|0|["']false["']|["']0["'])/gi, remediation: 'Enable SSL/TLS certificate validation in production', references: ['https://owasp.org/Top10/A05_2021-Security_Misconfiguration/'] }, { id: 'OWASP-A07-001', name: 'Weak Authentication', description: 'Potential weak authentication mechanism', severity: 'high', category: 'code', pattern: /(?:md5|sha1)\s*\(/gi, remediation: 'Use strong hashing algorithms like bcrypt, scrypt, or Argon2 for passwords', references: ['https://owasp.org/Top10/A07_2021-Identification_and_Authentication_Failures/'] } ]; } /** * CWE Top 25 security rules */ getCWERules(): SecurityRule[] { return [ { id: 'CWE-79-001', name: 'Reflected XSS', description: 'User input reflected without encoding', severity: 'high', category: 'code', pattern: /res\.(?:send|write|end)\s*\([^)]*(?:req\.(?:query|params|body)|request\.)/g, remediation: 'Encode all user input before reflecting in responses', references: ['https://cwe.mitre.org/data/definitions/79.html'] }, { id: 'CWE-89-001', name: 'SQL String Concatenation', description: 'SQL query built using string concatenation', severity: 'critical', category: 'code', pattern: /["'`].*(?:SELECT|INSERT|UPDATE|DELETE|DROP|CREATE|ALTER).*["'`]\s*\+\s*\w+/gi, remediation: 'Use parameterized queries instead of string concatenation', references: ['https://cwe.mitre.org/data/definitions/89.html'] }, { id: 'CWE-22-001', name: 'Path Manipulation', description: 'File path constructed from user input', severity: 'high', category: 'code', pattern: /path\.join\s*\([^)]*(?:req\.|request\.|params|query|body)/g, remediation: 'Validate paths against a whitelist and use path.resolve()', references: ['https://cwe.mitre.org/data/definitions/22.html'] }, { id: 'CWE-798-001', name: 'Hardcoded Credentials', description: 'Credentials hardcoded in source', severity: 'critical', category: 'code', pattern: /(?:username|user|login)\s*[:=]\s*["'][^"']+["'].*(?:password|pass|pwd)\s*[:=]\s*["'][^"']+["']/gi, remediation: 'Store credentials in environment variables or secure vaults', references: ['https://cwe.mitre.org/data/definitions/798.html'] } ]; } /** * DollhouseMCP-specific security rules */ getDollhouseMCPRules(): SecurityRule[] { return [ { id: 'DMCP-SEC-001', name: 'Unsafe Persona Loading', description: 'Persona loaded without validation', severity: 'high', category: 'custom', pattern: /loadPersona\s*\([^)]*\)\s*(?!.*validate)/g, remediation: 'Always validate personas before loading using PersonaValidator', references: ['DollhouseMCP Security Guidelines'] }, { id: 'DMCP-SEC-002', name: 'Token Validation Bypass', description: 'Token used without validation', severity: 'critical', category: 'custom', pattern: /(?:getToken|useToken|token\.use)\s*\([^)]*\)(?!.*(?:validate|verify|check))/gi, remediation: 'Always validate tokens using TokenManager.validateToken()', references: ['DollhouseMCP Security Guidelines'] }, { id: 'DMCP-SEC-003', name: 'Rate Limiting Missing', description: 'API endpoint without rate limiting', severity: 'medium', category: 'custom', check: (content, context) => { const findings: SecurityFinding[] = []; // Check for MCP tool handlers without rate limiting const toolPattern = /name:\s*["']([^"']+)["'].*handle:/gs; const hasRateLimit = /rateLimiter|checkRateLimit|tokenBucket/i.test(content); if (toolPattern.test(content) && !hasRateLimit) { findings.push({ ruleId: 'DMCP-SEC-003', severity: 'medium' as const, message: 'MCP tool handler without rate limiting', remediation: 'Add rate limiting to prevent abuse', confidence: 'high' as const }); } return findings; }, remediation: 'Implement rate limiting for all MCP tools', references: ['Issue #174 - Rate Limiting Implementation'] }, { id: 'DMCP-SEC-004', name: 'Unicode Validation Missing', description: 'User input processed without Unicode normalization', severity: 'medium', category: 'custom', check: (content, context) => { const findings: SecurityFinding[] = []; // Check for user input processing without Unicode validation const inputPattern = /(?:req\.|request\.|params|query|body|content)/; const hasUnicodeCheck = /UnicodeValidator|normalizeUnicode/i.test(content); if (inputPattern.test(content) && !hasUnicodeCheck) { findings.push({ ruleId: 'DMCP-SEC-004', severity: 'medium' as const, message: 'User input processed without Unicode normalization', remediation: 'Use UnicodeValidator.normalize() on all user input', confidence: 'medium' as const }); } return findings; }, remediation: 'Apply Unicode normalization to prevent bypass attacks', references: ['Issue #162 - Unicode Normalization'] }, { id: 'DMCP-SEC-005', name: 'Unvalidated YAML Content', description: 'YAML content parsed without security validation', severity: 'high', category: 'custom', pattern: /yaml\.load\s*\(|parse\s*\([^)]*\.ya?ml/gi, remediation: 'Use SecureYamlParser for all YAML parsing', references: ['DollhouseMCP Security Guidelines'] }, { id: 'DMCP-SEC-006', name: 'Security Event Not Logged', description: 'Security-relevant operation without logging', severity: 'low', category: 'custom', check: (content, context) => { const findings: SecurityFinding[] = []; // Check for security operations without logging const securityOps = /(?:authenticate|authorize|validate|sanitize|encrypt|decrypt)/i; const hasLogging = /SecurityMonitor\.log|logSecurityEvent/i.test(content); if (securityOps.test(content) && !hasLogging) { findings.push({ ruleId: 'DMCP-SEC-006', severity: 'low' as const, message: 'Security operation without audit logging', remediation: 'Add SecurityMonitor.logSecurityEvent() for audit trail', confidence: 'medium' as const }); } return findings; }, remediation: 'Log all security-relevant operations for audit trail', references: ['DollhouseMCP Security Guidelines'] } ]; } }