@bryan-thompson/inspector-assessment-client 1.26.4 → 1.26.6

package/lib/services/assessment/modules/securityTests/SecurityResponseAnalyzer.js

@@ -1,16 +1,48 @@
 /**
- * Security Response Analyzer
+ * Security Response Analyzer (Facade)
  * Analyzes tool responses for evidence-based vulnerability detection
  *
- * Extracted from SecurityAssessor.ts for maintainability.
- * Handles response analysis, reflection detection, and confidence calculation.
+ * REFACTORED in Issue #53 (v2.0.0): Converted to facade pattern
+ * Delegates to focused classes for maintainability (CC 218 → ~50)
+ *
+ * Extracted classes:
+ * - ErrorClassifier: Error classification and connection error detection
+ * - ExecutionArtifactDetector: Execution evidence detection
+ * - MathAnalyzer: Math computation detection (Calculator Injection)
+ * - SafeResponseDetector: Safe response pattern detection
+ * - ConfidenceScorer: Confidence level calculation
  */
 import { ToolClassifier, ToolCategory } from "../../ToolClassifier.js";
+// Import extracted classes
+import { ErrorClassifier } from "./ErrorClassifier.js";
+import { ExecutionArtifactDetector } from "./ExecutionArtifactDetector.js";
+import { MathAnalyzer } from "./MathAnalyzer.js";
+import { SafeResponseDetector } from "./SafeResponseDetector.js";
+import { ConfidenceScorer } from "./ConfidenceScorer.js";
 /**
  * Analyzes tool responses for security vulnerabilities
  * Distinguishes between safe reflection and actual execution
+ *
+ * This class serves as a facade, delegating to focused analyzers
+ * while maintaining the same public API for backward compatibility.
  */
 export class SecurityResponseAnalyzer {
+    // Delegate classes
+    errorClassifier;
+    executionDetector;
+    mathAnalyzer;
+    safeDetector;
+    confidenceScorer;
+    constructor() {
+        this.errorClassifier = new ErrorClassifier();
+        this.executionDetector = new ExecutionArtifactDetector();
+        this.mathAnalyzer = new MathAnalyzer();
+        this.safeDetector = new SafeResponseDetector();
+        this.confidenceScorer = new ConfidenceScorer();
+    }
+    // ============================================================================
+    // PUBLIC API - These 8 methods maintain backward compatibility
+    // ============================================================================
     /**
      * Analyze response with evidence-based detection
      * CRITICAL: Distinguish between safe reflection and actual execution
@@ -20,7 +52,7 @@ export class SecurityResponseAnalyzer {
      */
     analyzeResponse(response, payload, tool) {
         const responseText = this.extractResponseContent(response).toLowerCase();
-        const errorInfo = this.extractErrorInfo(response);
+        const errorInfo = this.errorClassifier.extractErrorInfo(response);
         // Check 1: Safe error responses (MCP validation, HTTP errors)
         const errorResult = this.checkSafeErrorResponses(responseText, errorInfo);
         if (errorResult)
@@ -33,1265 +65,344 @@ export class SecurityResponseAnalyzer {
         return this.checkVulnerabilityEvidence(response, payload, responseText);
     }
     /**
-     * Check for safe error responses that indicate proper input rejection
-     * Handles: MCP validation errors (-32602), HTTP 4xx/5xx errors
+     * Calculate confidence level and manual review requirements
      */
-    checkSafeErrorResponses(responseText, errorInfo) {
-        // ✅ STEP 1: Check for MCP validation errors (HIGHEST PRIORITY)
-        // Tools that reject invalid input before processing are SECURE
-        if (this.isMCPValidationError(errorInfo, responseText)) {
-            const errorCode = errorInfo.code ? ` (${errorInfo.code})` : "";
-            return {
-                isVulnerable: false,
-                evidence: `MCP validation error${errorCode}: Tool properly rejected invalid input before processing`,
-            };
-        }
-        // ✅ STEP 1.1: Check for HTTP error responses (Issue #26)
-        // HTTP 4xx/5xx errors indicate tool rejection, not vulnerability
-        if (this.isHttpErrorResponse(responseText)) {
-            return {
-                isVulnerable: false,
-                evidence: "HTTP error response indicates tool rejection (not vulnerability)",
-            };
-        }
-        return null; // Continue to next check
+    calculateConfidence(tool, isVulnerable, evidence, responseText, payload, sanitizationResult) {
+        return this.confidenceScorer.calculateConfidence(tool, isVulnerable, evidence, responseText, payload, sanitizationResult);
     }
     /**
-     * Check for safe tool behavior patterns
-     * Handles: Tool categories, reflection, computed math, validation rejection
+     * Analyze response for auth bypass patterns (Issue #75)
+     * Detects fail-open authentication vulnerabilities (CVE-2025-52882)
      */
-    checkSafeToolBehavior(response, payload, tool, responseText) {
-        // ✅ STEP 1.5: Classify tool and check for safe categories (prevents false positives)
-        const classifier = new ToolClassifier();
-        const classification = classifier.classify(tool.name, tool.description);
-        // Check if tool is in a safe category and response matches expected format
-        if (classification.categories.includes(ToolCategory.SEARCH_RETRIEVAL)) {
-            if (this.isSearchResultResponse(responseText)) {
-                return {
-                    isVulnerable: false,
-                    evidence: "Search tool returned query results (not code execution)",
-                };
+    analyzeAuthBypassResponse(response) {
+        const responseText = this.extractResponseContent(response);
+        // Fail-open indicators (VULNERABLE - CVE-2025-52882 pattern)
+        const failOpenPatterns = [
+            {
+                pattern: /"auth_type"\s*:\s*"fail-open"/i,
+                evidence: "auth_type: fail-open (CVE-2025-52882)",
+            },
+            {
+                pattern: /"auth_status"\s*:\s*"bypassed"/i,
+                evidence: "auth_status: bypassed",
+            },
+            {
+                pattern: /access.*granted.*despite/i,
+                evidence: "access granted despite failure",
+            },
+            {
+                pattern: /authentication.*skipped/i,
+                evidence: "authentication skipped",
+            },
+            { pattern: /fail[\s-]?open/i, evidence: "fail-open pattern detected" },
+            { pattern: /bypassed.*auth/i, evidence: "auth bypassed" },
+            {
+                pattern: /authentication.*bypassed/i,
+                evidence: "authentication bypassed",
+            },
+            {
+                pattern: /"vulnerable"\s*:\s*true[^}]*(?:"auth_status"|"auth_type"|"auth_error")|(?:"auth_status"|"auth_type"|"auth_error")[^}]*"vulnerable"\s*:\s*true/i,
+                evidence: "vulnerable flag with auth context",
+            },
+            {
+                pattern: /auth.*succeeded.*null/i,
+                evidence: "auth succeeded with null token",
+            },
+            {
+                pattern: /granted.*without.*valid/i,
+                evidence: "granted without valid token",
+            },
+            {
+                pattern: /"action_performed"/i,
+                evidence: "action performed indicator",
+            },
+        ];
+        // Fail-closed indicators (SAFE - proper auth handling)
+        const failClosedPatterns = [
+            {
+                pattern: /"auth_type"\s*:\s*"fail-closed"/i,
+                evidence: "auth_type: fail-closed (secure)",
+            },
+            {
+                pattern: /"auth_status"\s*:\s*"denied"/i,
+                evidence: "auth_status: denied",
+            },
+            { pattern: /access.*denied/i, evidence: "access denied" },
+            { pattern: /authentication.*failed/i, evidence: "authentication failed" },
+            {
+                pattern: /fail[\s-]?closed/i,
+                evidence: "fail-closed pattern detected",
+            },
+            { pattern: /"status"\s*:\s*"blocked"/i, evidence: "status: blocked" },
+            { pattern: /invalid.*token/i, evidence: "invalid token rejection" },
+            { pattern: /token.*required/i, evidence: "token required" },
+            { pattern: /unauthorized/i, evidence: "unauthorized response" },
+            { pattern: /"denial_reason"/i, evidence: "denial reason provided" },
+        ];
+        // Check for fail-open (vulnerable) patterns first
+        for (const { pattern, evidence } of failOpenPatterns) {
+            if (pattern.test(responseText)) {
+                return { detected: true, failureMode: "FAIL_OPEN", evidence };
             }
         }
-        if (classification.categories.includes(ToolCategory.CRUD_CREATION)) {
-            if (this.isCreationResponse(responseText)) {
-                return {
-                    isVulnerable: false,
-                    evidence: "CRUD tool created/modified resource (not code execution)",
-                };
+        // Check for fail-closed (safe) patterns
+        for (const { pattern, evidence } of failClosedPatterns) {
+            if (pattern.test(responseText)) {
+                return { detected: false, failureMode: "FAIL_CLOSED", evidence };
             }
         }
-        if (classification.categories.includes(ToolCategory.READ_ONLY_INFO)) {
-            return {
-                isVulnerable: false,
-                evidence: "Read-only info tool returned user/workspace data (intended data exposure, not vulnerability)",
-            };
-        }
-        if (classification.categories.includes(ToolCategory.SAFE_STORAGE)) {
-            return {
-                isVulnerable: false,
-                evidence: "Safe storage control tool (validated safe implementation, control group for testing)",
-            };
-        }
-        // ✅ STEP 2: Check if response is just reflecting/echoing input (SAFE)
-        const isJustReflection = this.isReflectionResponse(responseText);
-        if (isJustReflection) {
-            return {
-                isVulnerable: false,
-                evidence: "Tool safely reflected input without execution",
-            };
-        }
-        // ✅ STEP 1.7: Check for computed math results (Issue #14 fix, enhanced in Issue #58)
-        // Use enhanced analysis with tool context and confidence levels
-        const mathAnalysis = this.analyzeComputedMathResult(payload.payload, responseText, tool);
-        // Only flag as vulnerable if HIGH confidence (Issue #58 fix)
-        // Low/medium confidence excluded to prevent false positives on API wrapper tools
-        if (mathAnalysis.isComputed && mathAnalysis.confidence === "high") {
-            return {
-                isVulnerable: true,
-                evidence: `Tool computed math expression result instead of storing/echoing it (payload: ${payload.payload})`,
-            };
-        }
-        // ✅ STEP 2: Check if tool explicitly rejected the input (SAFE)
-        if (this.isValidationRejection(response)) {
-            return {
-                isVulnerable: false,
-                evidence: "Tool correctly rejected invalid input with validation error",
-            };
-        }
-        return null; // Continue to next check
+        return { detected: false, failureMode: "UNKNOWN" };
     }
     /**
-     * Check for vulnerability evidence in response
-     * Handles: Evidence pattern matching, fallback injection analysis
+     * Check if response indicates connection/server failure
      */
-    checkVulnerabilityEvidence(response, payload, responseText) {
-        const payloadText = payload.payload.toLowerCase();
-        // ✅ STEP 3: Check for expected evidence of execution
-        if (payload.evidence && payload.evidence.test(responseText)) {
-            const evidenceMatchesInput = payload.evidence.test(payloadText);
-            if (evidenceMatchesInput && responseText.includes(payloadText)) {
-                return {
-                    isVulnerable: false,
-                    evidence: "Evidence pattern matched echoed input, not execution",
-                };
-            }
-            // For validation-type patterns, require execution evidence
-            if (this.isValidationPattern(payload.evidence)) {
-                const hasExecutionEvidence = this.hasExecutionEvidence(responseText);
-                if (!hasExecutionEvidence) {
-                    return {
-                        isVulnerable: false,
-                        evidence: "Pattern matched in error message but no execution evidence - tool properly validated input",
-                    };
-                }
-            }
-            return {
-                isVulnerable: true,
-                evidence: `Tool executed malicious instruction: found evidence matching ${payload.evidence} - ${payload.description}`,
-            };
-        }
-        // ✅ STEP 4: Fall back to existing detection logic
-        return this.analyzeInjectionResponse(response, payload.payload);
+    isConnectionError(response) {
+        return this.errorClassifier.isConnectionError(response);
     }
     /**
-     * Check if tool explicitly rejected input with validation error (SAFE)
+     * Check if caught exception indicates connection/server failure
      */
-    isValidationRejection(response) {
-        const responseText = this.extractResponseContent(response);
-        try {
-            const parsed = JSON.parse(responseText);
-            if (parsed.valid === false ||
-                parsed.error === true ||
-                parsed.error === "true" ||
-                (parsed.error && parsed.error !== false) ||
-                parsed.status === "rejected" ||
-                parsed.status === "invalid" ||
-                parsed.status === "failed") {
-                return true;
-            }
-            if (parsed.errors &&
-                Array.isArray(parsed.errors) &&
-                parsed.errors.length > 0) {
-                return true;
-            }
-            if (parsed.error && typeof parsed.error === "string") {
-                return true;
-            }
-            if (typeof parsed.result === "string") {
-                const resultRejectionPatterns = [
-                    /validation (failed|error)/i,
-                    /rejected/i,
-                    /not.*approved/i,
-                    /not.*in.*list/i,
-                    /invalid.*input/i,
-                    /error:.*invalid/i,
-                ];
-                if (resultRejectionPatterns.some((p) => p.test(parsed.result))) {
-                    return true;
-                }
-            }
-        }
-        catch {
-            // Not JSON, check text patterns
-        }
-        const rejectionPatterns = [
-            /validation failed/i,
-            /rejected/i,
-            /not.*approved/i,
-            /not.*in.*list/i,
-            /invalid.*input/i,
-            /error:.*invalid/i,
-        ];
-        return rejectionPatterns.some((pattern) => pattern.test(responseText));
+    isConnectionErrorFromException(error) {
+        return this.errorClassifier.isConnectionErrorFromException(error);
+    }
+    /**
+     * Classify error type for reporting
+     */
+    classifyError(response) {
+        return this.errorClassifier.classifyError(response);
+    }
+    /**
+     * Classify error type from caught exception
+     */
+    classifyErrorFromException(error) {
+        return this.errorClassifier.classifyErrorFromException(error);
+    }
+    /**
+     * Extract response content from MCP response
+     */
+    extractResponseContent(response) {
+        return this.safeDetector.extractResponseContent(response);
     }
+    // ============================================================================
+    // DELEGATED PUBLIC METHODS - Exposed for external use
+    // ============================================================================
     /**
      * Check if response is an MCP validation error (safe rejection)
      */
     isMCPValidationError(errorInfo, responseText) {
-        if (errorInfo.code === -32602 || errorInfo.code === "-32602") {
-            return true;
-        }
-        const validationPatterns = [
-            /parameter validation failed/i,
-            /schema validation (error|failed)/i,
-            /invalid (url|email|format|parameter|input|data)/i,
-            /must be a valid/i,
-            /must have a valid/i,
-            /failed to validate/i,
-            /validation error/i,
-            /does not match (pattern|schema)/i,
-            /not a valid (url|email|number|string)/i,
-            /expected.*but (got|received)/i,
-            /type mismatch/i,
-            /\brequired\b.*\bmissing\b/i,
-            /cannot.*be.*empty/i,
-            /must.*not.*be.*empty/i,
-            /empty.*not.*allowed/i,
-            /\brequired\b/i,
-            /missing.*required/i,
-            /field.*required/i,
-        ];
-        return validationPatterns.some((pattern) => pattern.test(responseText));
+        return this.safeDetector.isMCPValidationError(errorInfo, responseText);
     }
     /**
      * Check if response is an HTTP error (Issue #26)
      */
     isHttpErrorResponse(responseText) {
-        const httpErrorPattern = /\b(4\d{2}|5\d{2})\b.*?(not found|error|bad request|unauthorized|forbidden|internal server|unavailable|timeout|service)/i;
-        const simpleHttpPattern = /^(4\d{2}|5\d{2})[\s:]/;
-        const notFoundPattern = /not found/i;
-        const isShortNotFound = notFoundPattern.test(responseText) && responseText.length < 100;
-        const jsonStatusPattern = /"status":\s*(4\d{2}|5\d{2})/;
-        return (httpErrorPattern.test(responseText) ||
-            simpleHttpPattern.test(responseText) ||
-            isShortNotFound ||
-            jsonStatusPattern.test(responseText));
+        return this.safeDetector.isHttpErrorResponse(responseText);
     }
     /**
      * Check if evidence pattern is ambiguous
      */
     isValidationPattern(evidencePattern) {
-        const patternStr = evidencePattern.toString().toLowerCase();
-        const ambiguousPatterns = [
-            "type.*error",
-            "invalid.*type",
-            "error",
-            "invalid",
-            "failed",
-            "negative.*not.*allowed",
-            "must.*be.*positive",
-            "invalid.*value",
-            "overflow",
-            "out.*of.*range",
-        ];
-        return ambiguousPatterns.some((ambiguous) => patternStr.includes(ambiguous));
+        return this.confidenceScorer.isValidationPattern(evidencePattern);
     }
     /**
      * Check if response contains evidence of actual execution
      */
     hasExecutionEvidence(responseText) {
-        const executionIndicators = [
-            /\bexecuted\b/i,
-            /\bprocessed\b/i,
-            /\bran\b.*command/i,
-            /\bcompleted\b/i,
-            /\bcomputed\b/i,
-            /\bcalculated\b/i,
-            /NullPointerException/i,
-            /SegmentationFault/i,
-            /StackOverflow/i,
-            /OutOfMemory/i,
-            /syntax error in executed/i,
-            /error while executing/i,
-            /failed during execution/i,
-            /error in query execution/i,
-            /runtime error/i,
-            /deleted \d+ (rows|files|records)/i,
-            /(file|resource) (opened|accessed|modified|deleted)/i,
-            /query returned \d+ results/i,
-            /modified \d+ records/i,
-            /\d+ rows affected/i,
-            /command output:/i,
-            /execution result:/i,
-        ];
-        return executionIndicators.some((pattern) => pattern.test(responseText));
+        return this.executionDetector.hasExecutionEvidence(responseText);
     }
     /**
      * Check if a math expression payload was computed (execution evidence)
+     * @deprecated Use analyzeComputedMathResult instead
      */
     isComputedMathResult(payload, responseText) {
-        // Skip HTTP error responses
-        const httpErrorPattern = /\b(4\d{2}|5\d{2})\b.*?(not found|error|bad request|unauthorized|forbidden|internal server|unavailable|timeout)/i;
-        const simpleHttpPattern = /^(4\d{2}|5\d{2})[\s:]/;
-        const notFoundPattern = /not found/i;
-        if (httpErrorPattern.test(responseText) ||
-            simpleHttpPattern.test(responseText) ||
-            (notFoundPattern.test(responseText) && responseText.length < 100)) {
-            return false;
-        }
-        const simpleMathPattern = /^\s*(\d+)\s*([+\-*/])\s*(\d+)(?:\s*([+\-*/])\s*(\d+))?\s*$/;
-        const match = payload.match(simpleMathPattern);
-        if (!match) {
-            return false;
-        }
-        try {
-            const num1 = parseInt(match[1], 10);
-            const op1 = match[2];
-            const num2 = parseInt(match[3], 10);
-            const op2 = match[4];
-            const num3 = match[5] ? parseInt(match[5], 10) : undefined;
-            let result;
-            switch (op1) {
-                case "+":
-                    result = num1 + num2;
-                    break;
-                case "-":
-                    result = num1 - num2;
-                    break;
-                case "*":
-                    result = num1 * num2;
-                    break;
-                case "/":
-                    result = Math.floor(num1 / num2);
-                    break;
-                default:
-                    return false;
-            }
-            if (op2 && num3 !== undefined) {
-                switch (op2) {
-                    case "+":
-                        result = result + num3;
-                        break;
-                    case "-":
-                        result = result - num3;
-                        break;
-                    case "*":
-                        result = result * num3;
-                        break;
-                    case "/":
-                        result = Math.floor(result / num3);
-                        break;
-                    default:
-                        return false;
-                }
-            }
-            const resultStr = result.toString();
-            const hasComputedResult = responseText.includes(resultStr);
-            const normalizedPayload = payload.replace(/\s+/g, "");
-            const hasOriginalExpression = responseText.includes(payload) ||
-                responseText.includes(normalizedPayload);
-            return hasComputedResult && !hasOriginalExpression;
-        }
-        catch {
-            return false;
-        }
+        return this.mathAnalyzer.isComputedMathResult(payload, responseText);
     }
     /**
-     * Check if numeric value appears in structured data context (not as computation result)
-     * Distinguishes {"records": 4} from computed "4" (Issue #58)
-     *
-     * @param result The computed numeric result to check for
-     * @param responseText The response text to analyze
-     * @returns true if the number appears to be coincidental data, not a computed result
+     * Check if numeric value appears in structured data context
      */
     isCoincidentalNumericInStructuredData(result, responseText) {
-        // Common data field names that often contain numeric values
-        const dataFieldPatterns = [
-            "count",
-            "total",
-            "records",
-            "page",
-            "limit",
-            "offset",
-            "id",
-            "status",
-            "code",
-            "version",
-            "index",
-            "size",
-            "employees",
-            "items",
-            "results",
-            "entries",
-            "length",
-            "pages",
-            "rows",
-            "columns",
-            "width",
-            "height",
-            "timestamp",
-            "duration",
-            "amount",
-            "price",
-            "quantity",
-        ];
-        // Try to parse as JSON
-        try {
-            const parsed = JSON.parse(responseText);
-            const checkObject = (obj, depth = 0) => {
-                if (depth > 5)
-                    return false; // Prevent deep recursion
-                if (typeof obj !== "object" || obj === null)
-                    return false;
-                for (const [key, value] of Object.entries(obj)) {
-                    // Check if numeric value matches result and key is a data field
-                    if (value === result) {
-                        const keyLower = key.toLowerCase();
-                        if (dataFieldPatterns.some((pattern) => keyLower.includes(pattern))) {
-                            return true;
-                        }
-                    }
-                    // Recurse into nested objects
-                    if (typeof value === "object" && value !== null) {
-                        if (checkObject(value, depth + 1))
-                            return true;
-                    }
-                    // Check arrays
-                    if (Array.isArray(value)) {
-                        for (const item of value) {
-                            if (typeof item === "object" && checkObject(item, depth + 1)) {
-                                return true;
-                            }
-                        }
-                    }
-                }
-                return false;
-            };
-            return checkObject(parsed);
-        }
-        catch {
-            // Not JSON - check for structured text patterns
-            // e.g., "Records: 4" or "Page 1 of 4" or "Total: 4 items"
-            const structuredPatterns = [
-                new RegExp(`(records|count|total|page|items|results|employees|entries|rows)[:\\s]+${result}\\b`, "i"),
-                new RegExp(`\\b${result}\\s+(records|items|results|entries|employees|rows)\\b`, "i"),
-                new RegExp(`page\\s+\\d+\\s+of\\s+${result}\\b`, "i"),
-                new RegExp(`total[:\\s]+${result}\\b`, "i"),
-                new RegExp(`found\\s+${result}\\s+(results|items|entries)`, "i"),
-            ];
-            return structuredPatterns.some((pattern) => pattern.test(responseText));
-        }
+        return this.mathAnalyzer.isCoincidentalNumericInStructuredData(result, responseText);
     }
     /**
      * Enhanced computed math result analysis with tool context (Issue #58)
-     *
-     * Returns a confidence level indicating how likely this is a real Calculator Injection:
-     * - high: Strong evidence of computation (should flag as vulnerable)
-     * - medium: Ambiguous (excluded from vulnerability count per user decision)
-     * - low: Likely coincidental data (excluded from vulnerability count)
      */
     analyzeComputedMathResult(payload, responseText, tool) {
-        // Skip HTTP error responses
-        const httpErrorPattern = /\b(4\d{2}|5\d{2})\b.*?(not found|error|bad request|unauthorized|forbidden|internal server|unavailable|timeout)/i;
-        const simpleHttpPattern = /^(4\d{2}|5\d{2})[\s:]/;
-        const notFoundPattern = /not found/i;
-        if (httpErrorPattern.test(responseText) ||
-            simpleHttpPattern.test(responseText) ||
-            (notFoundPattern.test(responseText) && responseText.length < 100)) {
-            return {
-                isComputed: false,
-                confidence: "high",
-                reason: "HTTP error response",
-            };
-        }
-        const simpleMathPattern = /^\s*(\d+)\s*([+\-*/])\s*(\d+)(?:\s*([+\-*/])\s*(\d+))?\s*$/;
-        const match = payload.match(simpleMathPattern);
-        if (!match) {
-            return {
-                isComputed: false,
-                confidence: "high",
-                reason: "Not a math expression",
-            };
-        }
-        try {
-            const num1 = parseInt(match[1], 10);
-            const op1 = match[2];
-            const num2 = parseInt(match[3], 10);
-            const op2 = match[4];
-            const num3 = match[5] ? parseInt(match[5], 10) : undefined;
-            let result;
-            switch (op1) {
-                case "+":
-                    result = num1 + num2;
-                    break;
-                case "-":
-                    result = num1 - num2;
-                    break;
-                case "*":
-                    result = num1 * num2;
-                    break;
-                case "/":
-                    result = Math.floor(num1 / num2);
-                    break;
-                default:
-                    return {
-                        isComputed: false,
-                        confidence: "high",
-                        reason: "Invalid operator",
-                    };
-            }
-            if (op2 && num3 !== undefined) {
-                switch (op2) {
-                    case "+":
-                        result = result + num3;
-                        break;
-                    case "-":
-                        result = result - num3;
-                        break;
-                    case "*":
-                        result = result * num3;
-                        break;
-                    case "/":
-                        result = Math.floor(result / num3);
-                        break;
-                    default:
-                        return {
-                            isComputed: false,
-                            confidence: "high",
-                            reason: "Invalid second operator",
-                        };
-                }
-            }
-            const resultStr = result.toString();
-            const hasComputedResult = responseText.includes(resultStr);
-            const normalizedPayload = payload.replace(/\s+/g, "");
-            const hasOriginalExpression = responseText.includes(payload) ||
-                responseText.includes(normalizedPayload);
-            // Basic detection: result present without original expression
-            const basicDetection = hasComputedResult && !hasOriginalExpression;
-            if (!basicDetection) {
-                return {
-                    isComputed: false,
-                    confidence: "high",
-                    reason: "No computed result found",
-                };
-            }
-            // Layer 1: Check if numeric appears in structured data context (Issue #58)
-            if (this.isCoincidentalNumericInStructuredData(result, responseText)) {
-                return {
-                    isComputed: false,
-                    confidence: "low",
-                    reason: "Numeric value appears in structured data field (e.g., count, records)",
-                };
-            }
-            // Layer 2: Tool classification heuristics (Issue #58)
-            if (tool) {
-                const classifier = new ToolClassifier();
-                const classification = classifier.classify(tool.name, tool.description);
-                // Check for read-only/data fetcher categories
-                if (classification.categories.includes(ToolCategory.DATA_FETCHER) ||
-                    classification.categories.includes(ToolCategory.API_WRAPPER) ||
-                    classification.categories.includes(ToolCategory.SEARCH_RETRIEVAL)) {
-                    return {
-                        isComputed: false,
-                        confidence: "low",
-                        reason: `Tool classified as ${classification.categories[0]} - unlikely to compute math`,
-                    };
-                }
-                // Check for "get_", "list_", "fetch_" patterns in tool name
-                const readOnlyNamePatterns = /^(get|list|fetch|read|retrieve|show|view)_/i;
-                if (readOnlyNamePatterns.test(tool.name)) {
-                    return {
-                        isComputed: false,
-                        confidence: "low",
-                        reason: "Tool name indicates read-only operation",
-                    };
-                }
-            }
-            // Layer 3: Check for computational language in response
-            const computationalIndicators = [
-                /\bthe\s+answer\s+is\b/i,
-                /\bresult\s*[=:]\s*\d/i,
-                /\bcalculated\s+to\b/i,
-                /\bcomputed\s+as\b/i,
-                /\bevaluates?\s+to\b/i,
-                /\bequals?\s+\d/i,
-                /\bsum\s+is\b/i,
-                /\bproduct\s+is\b/i,
-            ];
-            const hasComputationalContext = computationalIndicators.some((p) => p.test(responseText));
-            if (hasComputationalContext) {
-                return {
-                    isComputed: true,
-                    confidence: "high",
-                    reason: "Response contains computational language",
-                };
-            }
-            // Layer 4: Longer responses without computational language are likely data
-            if (responseText.length > 50) {
-                return {
-                    isComputed: false,
-                    confidence: "medium",
-                    reason: "Response lacks computational language, likely coincidental data",
-                };
-            }
-            // Short response with just the number - this is suspicious
-            if (responseText.trim() === resultStr) {
-                return {
-                    isComputed: true,
-                    confidence: "high",
-                    reason: "Response is exactly the computed result",
-                };
-            }
-            // Default: medium confidence (excluded per user decision)
-            return {
-                isComputed: false,
-                confidence: "medium",
-                reason: "Ambiguous - numeric match without computational context",
-            };
-        }
-        catch {
-            return { isComputed: false, confidence: "high", reason: "Parse error" };
-        }
+        return this.mathAnalyzer.analyzeComputedMathResult(payload, responseText, tool);
     }
     /**
-     * Check if response indicates connection/server failure
+     * Check if response is just reflection (safe)
      */
-    isConnectionError(response) {
-        const text = this.extractResponseContent(response).toLowerCase();
-        const unambiguousPatterns = [
-            /MCP error -32001/i,
-            /MCP error -32603/i,
-            /MCP error -32000/i,
-            /MCP error -32700/i,
-            /socket hang up/i,
-            /ECONNREFUSED/i,
-            /ETIMEDOUT/i,
-            /ERR_CONNECTION/i,
-            /fetch failed/i,
-            /connection reset/i,
-            /error POSTing to endpoint/i,
-            /error GETting.*endpoint/i,
-            /service unavailable/i,
-            /gateway timeout/i,
-            /unknown tool:/i,
-            /no such tool/i,
-        ];
-        if (unambiguousPatterns.some((pattern) => pattern.test(text))) {
-            return true;
-        }
-        const mcpPrefix = /^mcp error -\d+:/i.test(text);
-        if (mcpPrefix) {
-            const contextualPatterns = [
-                /bad request/i,
-                /unauthorized/i,
-                /forbidden/i,
-                /no valid session/i,
-                /session.*expired/i,
-                /internal server error/i,
-                /HTTP [45]\d\d/i,
-            ];
-            return contextualPatterns.some((pattern) => pattern.test(text));
-        }
-        return false;
+    isReflectionResponse(responseText) {
+        return this.safeDetector.isReflectionResponse(responseText);
     }
     /**
-     * Check if caught exception indicates connection/server failure
+     * Detect execution artifacts in response
      */
-    isConnectionErrorFromException(error) {
-        if (error instanceof Error) {
-            const message = error.message.toLowerCase();
-            const unambiguousPatterns = [
-                /MCP error -32001/i,
-                /MCP error -32603/i,
-                /MCP error -32000/i,
-                /MCP error -32700/i,
-                /socket hang up/i,
-                /ECONNREFUSED/i,
-                /ETIMEDOUT/i,
-                /network error/i,
-                /ERR_CONNECTION/i,
-                /fetch failed/i,
-                /connection reset/i,
-                /error POSTing to endpoint/i,
-                /error GETting/i,
-                /service unavailable/i,
-                /gateway timeout/i,
-                /unknown tool:/i,
-                /no such tool/i,
-            ];
-            if (unambiguousPatterns.some((pattern) => pattern.test(message))) {
-                return true;
-            }
-            const mcpPrefix = /^mcp error -\d+:/i.test(message);
-            if (mcpPrefix) {
-                const contextualPatterns = [
-                    /bad request/i,
-                    /unauthorized/i,
-                    /forbidden/i,
-                    /no valid session/i,
-                    /session.*expired/i,
-                    /internal server error/i,
-                    /HTTP [45]\d\d/i,
-                ];
-                return contextualPatterns.some((pattern) => pattern.test(message));
-            }
-        }
-        return false;
+    detectExecutionArtifacts(responseText) {
+        return this.executionDetector.detectExecutionArtifacts(responseText);
     }
     /**
-     * Classify error type for reporting
+     * Check if response contains echoed injection payload patterns
      */
-    classifyError(response) {
-        const text = this.extractResponseContent(response).toLowerCase();
-        if (/socket|ECONNREFUSED|ETIMEDOUT|network|fetch failed|connection reset/i.test(text)) {
-            return "connection";
-        }
-        if (/-32603|-32000|-32700|internal server error|service unavailable|gateway timeout|HTTP 5\d\d|error POSTing.*endpoint|error GETting.*endpoint|bad request|HTTP 400|unauthorized|forbidden|no valid session|session.*expired/i.test(text)) {
-            return "server";
-        }
-        if (/-32001/i.test(text)) {
-            return "protocol";
-        }
-        return "protocol";
+    containsEchoedInjectionPayload(responseText) {
+        return this.executionDetector.containsEchoedInjectionPayload(responseText);
     }
     /**
-     * Classify error type from caught exception
+     * Check if tool explicitly rejected input with validation error (SAFE)
      */
-    classifyErrorFromException(error) {
-        if (error instanceof Error) {
-            const message = error.message.toLowerCase();
-            if (/socket|ECONNREFUSED|ETIMEDOUT|network|fetch failed|connection reset/i.test(message)) {
-                return "connection";
-            }
-            if (/-32603|-32000|-32700|internal server error|service unavailable|gateway timeout|HTTP 5\d\d|error POSTing|error GETting|bad request|HTTP 400|unauthorized|forbidden|no valid session|session.*expired/i.test(message)) {
-                return "server";
-            }
-            if (/-32001/i.test(message)) {
-                return "protocol";
-            }
-        }
-        return "protocol";
+    isValidationRejection(response) {
+        return this.safeDetector.isValidationRejection(response);
     }
     /**
-     * Check if response is just reflection (safe)
-     * Two-layer defense: Match reflection patterns, verify NO execution evidence
+     * Check if tool is a structured data tool
      */
-    isReflectionResponse(responseText) {
-        const statusPatterns = [
-            /\d+\s+total\s+(in\s+)?(memory|storage|items|results)/i,
-            /\d+\s+(results|items|records),?\s+\d+\s+total/i,
-            /action\s+executed\s+successfully:/i,
-            /command\s+executed\s+successfully:/i,
-            /"result":\s*"action\s+executed\s+successfully"/i,
-            /result.*action\s+executed\s+successfully/i,
-            /successfully\s+(executed|completed|processed):/i,
-            /successfully\s+(executed|completed|processed)"/i,
-            /action\s+received:/i,
-            /input\s+received:/i,
-            /request\s+received:/i,
-            /"safe"\s*:\s*true[^}]{0,500}("message"|"result"|"status"|"response")/i,
-            /("message"|"result"|"status"|"response")[^}]{0,500}"safe"\s*:\s*true/i,
-            /"vulnerable"\s*:\s*false[^}]{0,500}("safe"|"stored"|"reflected"|"status")/i,
-            /("safe"|"stored"|"reflected"|"status")[^}]{0,500}"vulnerable"\s*:\s*false/i,
-            /"status"\s*:\s*"acknowledged"[^}]{0,500}("message"|"result"|"safe")/i,
-            /("message"|"result"|"safe")[^}]{0,500}"status"\s*:\s*"acknowledged"/i,
-        ];
-        const reflectionPatterns = [
-            ...statusPatterns,
-            /stored.*query/i,
-            /saved.*input/i,
-            /received.*parameter/i,
-            /processing.*request/i,
-            /storing.*data/i,
-            /added.*to.*collection/i,
-            /echo:/i,
-            /echoing/i,
-            /repeating/i,
-            /displaying/i,
-            /showing.*input/i,
-            /message.*echoed/i,
-            /safely.*as.*data/i,
-            /query.*stored/i,
-            /input.*saved/i,
-            /parameter.*received/i,
-            /command.*stored/i,
-            /stored.*command/i,
-            /data.*stored/i,
-            /stored.*data/i,
-            /action.*stored/i,
-            /stored.*action/i,
-            /text.*stored/i,
-            /stored.*text/i,
-            /setting.*stored/i,
-            /stored.*setting/i,
-            /instruction.*stored/i,
-            /stored.*instruction/i,
-            /url.*stored/i,
-            /stored.*url/i,
-            /package.*stored/i,
-            /stored.*package/i,
-            /stored.*safely/i,
-            /safely.*stored/i,
-            /without\s+execut/i,
-            /not\s+executed/i,
-            /never\s+executed/i,
-            /stored.*as.*data/i,
-            /treated.*as.*data/i,
-            /stored\s+in\s+(collection|database)/i,
-            /stored.*successfully/i,
-            /saved.*to/i,
-            /recorded\s+in/i,
-            /added\s+to/i,
-            /logged successfully:/i,
-            /queued for processing:/i,
-            /saved (for|successfully)/i,
-            /stored for (admin review|configuration|processing)/i,
-            /processed successfully/i,
-            /validated successfully/i,
-            /parsed successfully/i,
-            /(validation|processing) (passed|completed)/i,
-            /error:.*not (found|in approved list|recognized)/i,
-            /error getting info for ['"].*['"]/i,
-            /invalid .* format.*stored as text/i,
-            /error:.*too (long|short|large)/i,
-            /payload.?rejected/i,
-            /input.?exceeds.?limit/i,
-            /resource.?limit.?enforced/i,
-            /size.?limit/i,
-            /maximum.?length/i,
-            /rate.?limit/i,
-            /request.?throttled/i,
-            /input.?too.?large/i,
-            /exceeds.?maximum.?size/i,
-            /depth.?limit.?exceeded/i,
-            /nesting.?limit/i,
-            /complexity.?limit/i,
-            /serialization.?not.?supported/i,
-            /pickle.?disabled/i,
-            /deserialization.?blocked/i,
-            /unsafe.?format.?rejected/i,
-            /binary.?data.?not.?accepted/i,
-            /data.?stored.?safely/i,
-            /without.?deserialization/i,
-            /no.?pickle/i,
-            /stored.?without.?deserializ/i,
-            /\[ref-[a-f0-9]+\]/i,
-            /stored.*\[ref-/i,
-            /\[sanitized\]/i,
-            /\[redacted\]/i,
-            /\[filtered\]/i,
-            /\[blocked\]/i,
-            /expression.*stored:/i,
-            /input.*sanitized/i,
-            /content.*replaced/i,
-        ];
-        const hasReflection = reflectionPatterns.some((pattern) => pattern.test(responseText));
-        if (hasReflection) {
-            try {
-                const parsed = JSON.parse(responseText);
-                const resultText = String(parsed.result || "");
-                const outputFields = [
-                    parsed.stdout,
-                    parsed.stderr,
-                    parsed.output,
-                    parsed.contents,
-                    parsed.execution_log,
-                    parsed.command_output,
-                ]
-                    .filter(Boolean)
-                    .join(" ");
-                const resultIsStatusOnly = statusPatterns.some((pattern) => pattern.test(resultText));
-                const hasExecutionInOutput = resultIsStatusOnly
-                    ? this.detectExecutionArtifacts(outputFields)
-                    : this.detectExecutionArtifacts(resultText) ||
-                        this.detectExecutionArtifacts(outputFields);
-                if (hasExecutionInOutput) {
-                    return false;
-                }
-                return true;
-            }
-            catch {
-                const hasExecution = this.detectExecutionArtifacts(responseText);
-                if (hasExecution) {
-                    return false;
-                }
-                return true;
-            }
-        }
-        try {
-            const parsed = JSON.parse(responseText);
-            if (parsed.action === "test" || parsed.action === "placeholder") {
-                const resultText = String(parsed.result || "");
-                if (!this.detectExecutionArtifacts(resultText)) {
-                    return true;
-                }
-            }
-            if (parsed.status && /(completed|success|ok|done)/.test(parsed.status)) {
-                if (!this.detectExecutionArtifacts(responseText)) {
-                    return true;
-                }
-            }
-        }
-        catch {
-            // Not JSON
-        }
-        return false;
+    isStructuredDataTool(toolName, toolDescription) {
+        return this.confidenceScorer.isStructuredDataTool(toolName, toolDescription);
     }
     /**
-     * Detect execution artifacts in response
+     * Check if response is returning search results
      */
-    detectExecutionArtifacts(responseText) {
-        const containsEchoedPayload = this.containsEchoedInjectionPayload(responseText);
-        const executionIndicators = [
-            /[a-z]+:x:\d+:\d+:/i,
-            /uid=\d+\([^)]+\)\s+gid=\d+/i,
-            /[d-][rwx-]{9}\s+\d+\s+[a-z]+/i,
-            /total\s+\d+\s*$/m,
-            /command_executed:\s*[^"\s]/i,
-            /stdout:\s*["']?[^"'\s]/i,
-            /(execution|output)_log:/i,
-            /\/bin\/(bash|sh|zsh|dash)/i,
-            /\b(root|administrator)\s*$/im,
-            /\/root\//i,
-            /PID:\s*\d{3,}/i,
-        ];
-        const contextSensitiveIndicators = [
-            /\/etc\/passwd/i,
-            /\/etc\/shadow/i,
-            /file:\/\/\//i,
-        ];
-        if (executionIndicators.some((pattern) => pattern.test(responseText))) {
-            return true;
-        }
-        if (!containsEchoedPayload) {
-            if (contextSensitiveIndicators.some((pattern) => pattern.test(responseText))) {
-                return true;
-            }
-        }
-        return false;
+    isSearchResultResponse(responseText) {
+        return this.safeDetector.isSearchResultResponse(responseText);
     }
     /**
-     * Check if response contains echoed injection payload patterns
+     * Check if response is from a creation/modification operation
      */
-    containsEchoedInjectionPayload(responseText) {
-        const echoedPayloadPatterns = [
-            /<!DOCTYPE\s+\w+\s+\[/i,
-            /<!ENTITY\s+\w+\s+SYSTEM/i,
-            /<!ENTITY\s+%\s*\w+/i,
-            /stored.*http:\/\//i,
-            /saved.*http:\/\//i,
-            /stored.*union\s+select/i,
-            /stored.*drop\s+table/i,
-            /stored\s+query:\s*[<'"]/i,
-            /saved\s+data:\s*[<'"]/i,
-        ];
-        return echoedPayloadPatterns.some((pattern) => pattern.test(responseText));
+    isCreationResponse(responseText) {
+        return this.safeDetector.isCreationResponse(responseText);
     }
+    // ============================================================================
+    // PRIVATE HELPER METHODS - Internal logic kept in facade
+    // ============================================================================
     /**
-     * Analyze injection response (fallback logic)
+     * Check for safe error responses that indicate proper input rejection
+     * Handles: MCP validation errors (-32602), HTTP 4xx/5xx errors
      */
-    analyzeInjectionResponse(response, _payload) {
-        const responseText = this.extractResponseContent(response);
-        const executionPatterns = [
-            /executed/i,
-            /command.*ran/i,
-            /result.*is/i,
-            /output.*:/i,
-            /returned.*value/i,
-        ];
-        const hasExecution = executionPatterns.some((pattern) => pattern.test(responseText));
-        if (hasExecution && !this.isReflectionResponse(responseText)) {
+    checkSafeErrorResponses(responseText, errorInfo) {
+        // MCP validation errors (HIGHEST PRIORITY)
+        if (this.safeDetector.isMCPValidationError(errorInfo, responseText)) {
+            const errorCode = errorInfo.code ? ` (${errorInfo.code})` : "";
             return {
-                isVulnerable: true,
-                evidence: "Tool executed instruction: found execution keywords",
+                isVulnerable: false,
+                evidence: `MCP validation error${errorCode}: Tool properly rejected invalid input before processing`,
             };
         }
-        return { isVulnerable: false };
+        // HTTP error responses (Issue #26)
+        if (this.safeDetector.isHttpErrorResponse(responseText)) {
+            return {
+                isVulnerable: false,
+                evidence: "HTTP error response indicates tool rejection (not vulnerability)",
+            };
+        }
+        return null;
     }
     /**
-     * Calculate confidence level and manual review requirements
-     *
-     * @param tool - The tool being tested
-     * @param isVulnerable - Whether the tool was flagged as vulnerable
-     * @param evidence - Evidence string from vulnerability detection
-     * @param responseText - The response text from the tool
-     * @param payload - The security payload used for testing
-     * @param sanitizationResult - Optional sanitization detection result (Issue #56)
-     * @returns Confidence result with manual review requirements
+     * Check for safe tool behavior patterns
+     * Handles: Tool categories, reflection, computed math, validation rejection
      */
-    calculateConfidence(tool, isVulnerable, evidence, responseText, payload, sanitizationResult) {
-        // Issue #56: If sanitization is detected, reduce confidence for vulnerabilities
-        // This helps reduce false positives on well-protected servers
-        if (isVulnerable && sanitizationResult?.detected) {
-            const adjustment = sanitizationResult.totalConfidenceAdjustment;
-            // Strong sanitization evidence (adjustment >= 30) - downgrade to low confidence
-            // This indicates the tool has specific security libraries in place
-            if (adjustment >= 30) {
-                const libraries = sanitizationResult.libraries.join(", ") || "general";
+    checkSafeToolBehavior(response, payload, tool, responseText) {
+        // Classify tool and check for safe categories
+        const classifier = new ToolClassifier();
+        const classification = classifier.classify(tool.name, tool.description);
+        // Check if tool is in a safe category
+        if (classification.categories.includes(ToolCategory.SEARCH_RETRIEVAL)) {
+            if (this.safeDetector.isSearchResultResponse(responseText)) {
                 return {
-                    confidence: "low",
-                    requiresManualReview: true,
-                    manualReviewReason: `Sanitization detected (${libraries}). ` +
-                        `Pattern match may be false positive due to security measures in place.`,
-                    reviewGuidance: `Tool uses sanitization libraries. Verify if the detected vulnerability ` +
-                        `actually bypasses the sanitization layer. Check: 1) Does the payload execute ` +
-                        `after sanitization? 2) Is the sanitization comprehensive for this attack type? ` +
-                        `3) Evidence: ${sanitizationResult.evidence.join("; ")}`,
+                    isVulnerable: false,
+                    evidence: "Search tool returned query results (not code execution)",
                 };
             }
-            // Moderate sanitization evidence (adjustment >= 15) - downgrade high to medium
-            if (adjustment >= 15) {
-                const patterns = sanitizationResult.libraries.length > 0
-                    ? sanitizationResult.libraries.join(", ")
-                    : sanitizationResult.genericPatterns.join(", ");
+        }
+        if (classification.categories.includes(ToolCategory.CRUD_CREATION)) {
+            if (this.safeDetector.isCreationResponse(responseText)) {
                 return {
-                    confidence: "medium",
-                    requiresManualReview: true,
-                    manualReviewReason: `Sanitization patterns detected (${patterns}). Verify actual vulnerability.`,
-                    reviewGuidance: `Tool mentions sanitization in description or shows sanitization in response. ` +
-                        `Verify if the detected pattern represents actual code execution or if it's ` +
-                        `safely handled. Evidence: ${sanitizationResult.evidence.join("; ")}`,
+                    isVulnerable: false,
+                    evidence: "CRUD tool created/modified resource (not code execution)",
                 };
             }
         }
-        const toolDescription = (tool.description || "").toLowerCase();
-        const toolName = tool.name.toLowerCase();
-        const responseLower = responseText.toLowerCase();
-        const payloadLower = payload.payload.toLowerCase();
-        // HIGH CONFIDENCE: Clear cases
-        if (!isVulnerable &&
-            (evidence.includes("safely reflected") ||
-                evidence.includes("API wrapper") ||
-                evidence.includes("safe: true"))) {
+        if (classification.categories.includes(ToolCategory.READ_ONLY_INFO)) {
             return {
-                confidence: "high",
-                requiresManualReview: false,
+                isVulnerable: false,
+                evidence: "Read-only info tool returned user/workspace data (intended data exposure, not vulnerability)",
             };
         }
-        if (isVulnerable &&
-            evidence.includes("executed") &&
-            !this.isStructuredDataTool(toolName, toolDescription)) {
+        if (classification.categories.includes(ToolCategory.SAFE_STORAGE)) {
             return {
-                confidence: "high",
-                requiresManualReview: false,
+                isVulnerable: false,
+                evidence: "Safe storage control tool (validated safe implementation, control group for testing)",
             };
         }
-        // LOW CONFIDENCE: Ambiguous pattern matches in structured data
-        if (isVulnerable) {
-            const isDataTool = this.isStructuredDataTool(toolName, toolDescription);
-            const hasStructuredData = /title:|name:|description:|trust score:|id:|snippets:/i.test(responseText) ||
-                /^\s*-\s+/m.test(responseText) ||
-                /"[^"]+"\s*:\s*"[^"]+"/g.test(responseText);
-            const patternInInput = payload.evidence?.test(payloadLower);
-            const echosInput = responseLower.includes(payloadLower);
-            if (isDataTool && (hasStructuredData || echosInput) && patternInInput) {
-                return {
-                    confidence: "low",
-                    requiresManualReview: true,
-                    manualReviewReason: "Pattern matched in structured data response. Tool may be legitimately " +
-                        "returning data containing search terms rather than executing malicious code.",
-                    reviewGuidance: "Verify: 1) Does the tool actually execute/compute the input? " +
-                        "2) Or does it just return pre-existing data that happens to contain the pattern? " +
-                        `3) Check if '${payload.evidence}' appears in legitimate tool output vs. execution results.`,
-                };
-            }
-            if (payload.evidence &&
-                /\b\d\b/.test(payload.evidence.toString()) &&
-                /\b(score|count|trust|rating|id|version)\b/i.test(responseText)) {
-                return {
-                    confidence: "low",
-                    requiresManualReview: true,
-                    manualReviewReason: "Numeric pattern found in response with numeric metadata (scores, counts, etc.). " +
-                        "May be coincidental data rather than arithmetic execution.",
-                    reviewGuidance: "Verify: 1) Did the tool actually compute an arithmetic result? " +
-                        "2) Or does the number appear in metadata like trust scores, version numbers, or counts? " +
-                        "3) Compare pattern location in response with tool's expected output format.",
-                };
-            }
-            if (/admin|role|privilege|elevated/i.test(payload.payload) &&
-                /\b(library|search|documentation|api|wrapper)\b/i.test(toolDescription)) {
-                return {
-                    confidence: "low",
-                    requiresManualReview: true,
-                    manualReviewReason: "Admin-related keywords found in search/retrieval tool results. " +
-                        "Tool may be returning data about admin-related libraries/APIs rather than elevating privileges.",
-                    reviewGuidance: "Verify: 1) Did the tool actually change behavior or assume admin role? " +
-                        "2) Or did it return search results for admin-related content? " +
-                        "3) Test if tool behavior actually changed after this request.",
-                };
-            }
+        // Check if response is just reflecting/echoing input (SAFE)
+        if (this.safeDetector.isReflectionResponse(responseText)) {
+            return {
+                isVulnerable: false,
+                evidence: "Tool safely reflected input without execution",
+            };
         }
-        // MEDIUM CONFIDENCE: Execution evidence but some ambiguity
-        if (isVulnerable && evidence.includes("executed")) {
+        // Check for computed math results (Issue #14 fix, enhanced in Issue #58)
+        const mathAnalysis = this.mathAnalyzer.analyzeComputedMathResult(payload.payload, responseText, tool);
+        // Only flag as vulnerable if HIGH confidence (Issue #58 fix)
+        if (mathAnalysis.isComputed && mathAnalysis.confidence === "high") {
             return {
-                confidence: "medium",
-                requiresManualReview: true,
-                manualReviewReason: "Execution indicators found but context suggests possible ambiguity.",
-                reviewGuidance: "Verify: 1) Review the full response to confirm actual code execution. " +
-                    "2) Check if tool's intended function involves execution. " +
-                    "3) Test with variations to confirm consistency.",
+                isVulnerable: true,
+                evidence: `Tool computed math expression result instead of storing/echoing it (payload: ${payload.payload})`,
             };
         }
-        // Default: HIGH confidence for clear safe cases
-        return {
-            confidence: "high",
-            requiresManualReview: false,
-        };
-    }
-    /**
-     * Check if tool is a structured data tool
-     */
-    isStructuredDataTool(toolName, toolDescription) {
-        const dataToolPatterns = [
-            /search/i,
-            /find/i,
-            /lookup/i,
-            /query/i,
-            /retrieve/i,
-            /fetch/i,
-            /get/i,
-            /list/i,
-            /resolve/i,
-            /discover/i,
-            /browse/i,
-        ];
-        const combined = `${toolName} ${toolDescription}`;
-        return dataToolPatterns.some((pattern) => pattern.test(combined));
-    }
-    /**
-     * Check if response is returning search results
-     */
-    isSearchResultResponse(responseText) {
-        const searchResultPatterns = [
-            /"results"\s*:\s*\[/i,
-            /"type"\s*:\s*"search"/i,
-            /"object"\s*:\s*"list"/i,
-            /\bhighlight\b/i,
-            /search\s+results/i,
-            /found\s+\d+\s+(results?|pages?|items?)/i,
-            /query\s+(returned|matched)/i,
-            /\d+\s+(results?|matches?|hits?)\s+for/i,
-            /"has_more"\s*:/i,
-            /next_cursor/i,
-        ];
-        return searchResultPatterns.some((pattern) => pattern.test(responseText));
+        // Check if tool explicitly rejected the input (SAFE)
+        if (this.safeDetector.isValidationRejection(response)) {
+            return {
+                isVulnerable: false,
+                evidence: "Tool correctly rejected invalid input with validation error",
+            };
+        }
+        return null;
     }
     /**
-     * Analyze response for auth bypass patterns (Issue #75)
-     * Detects fail-open authentication vulnerabilities (CVE-2025-52882)
-     *
-     * @param response The tool response to analyze
-     * @returns AuthBypassResult with detection status and failure mode classification
+     * Check for vulnerability evidence in response
+     * Handles: Evidence pattern matching, fallback injection analysis
      */
-    analyzeAuthBypassResponse(response) {
-        const responseText = this.extractResponseContent(response);
-        // Fail-open indicators (VULNERABLE - CVE-2025-52882 pattern)
-        const failOpenPatterns = [
-            {
-                pattern: /"auth_status"\s*:\s*"bypassed"/i,
-                evidence: "auth_status: bypassed",
-            },
-            {
-                pattern: /access.*granted.*despite/i,
-                evidence: "access granted despite failure",
-            },
-            {
-                pattern: /authentication.*skipped/i,
-                evidence: "authentication skipped",
-            },
-            { pattern: /fail[\s-]?open/i, evidence: "fail-open pattern detected" },
-            { pattern: /bypassed.*auth/i, evidence: "auth bypassed" },
-            {
-                pattern: /authentication.*bypassed/i,
-                evidence: "authentication bypassed",
-            },
-            {
-                pattern: /"vulnerable"\s*:\s*true/i,
-                evidence: "vulnerable: true flag",
-            },
-            {
-                pattern: /auth.*succeeded.*null/i,
-                evidence: "auth succeeded with null token",
-            },
-            {
-                pattern: /granted.*without.*valid/i,
-                evidence: "granted without valid token",
-            },
-            {
-                pattern: /"action_performed"/i,
-                evidence: "action performed indicator",
-            },
-        ];
-        // Fail-closed indicators (SAFE - proper auth handling)
-        const failClosedPatterns = [
-            {
-                pattern: /"auth_status"\s*:\s*"denied"/i,
-                evidence: "auth_status: denied",
-            },
-            { pattern: /access.*denied/i, evidence: "access denied" },
-            { pattern: /authentication.*failed/i, evidence: "authentication failed" },
-            {
-                pattern: /fail[\s-]?closed/i,
-                evidence: "fail-closed pattern detected",
-            },
-            { pattern: /"status"\s*:\s*"blocked"/i, evidence: "status: blocked" },
-            {
-                pattern: /invalid.*token/i,
-                evidence: "invalid token rejection",
-            },
-            {
-                pattern: /token.*required/i,
-                evidence: "token required",
-            },
-            {
-                pattern: /unauthorized/i,
-                evidence: "unauthorized response",
-            },
-            {
-                pattern: /"denial_reason"/i,
-                evidence: "denial reason provided",
-            },
-        ];
-        // Check for fail-open (vulnerable) patterns first
-        for (const { pattern, evidence } of failOpenPatterns) {
-            if (pattern.test(responseText)) {
-                return { detected: true, failureMode: "FAIL_OPEN", evidence };
+    checkVulnerabilityEvidence(response, payload, responseText) {
+        const payloadText = payload.payload.toLowerCase();
+        // Check for expected evidence of execution
+        if (payload.evidence && payload.evidence.test(responseText)) {
+            const evidenceMatchesInput = payload.evidence.test(payloadText);
+            if (evidenceMatchesInput && responseText.includes(payloadText)) {
+                return {
+                    isVulnerable: false,
+                    evidence: "Evidence pattern matched echoed input, not execution",
+                };
             }
-        }
-        // Check for fail-closed (safe) patterns
-        for (const { pattern, evidence } of failClosedPatterns) {
-            if (pattern.test(responseText)) {
-                return { detected: false, failureMode: "FAIL_CLOSED", evidence };
+            // For validation-type patterns, require execution evidence
+            if (this.confidenceScorer.isValidationPattern(payload.evidence)) {
+                const hasExecutionEvidence = this.executionDetector.hasExecutionEvidence(responseText);
+                if (!hasExecutionEvidence) {
+                    return {
+                        isVulnerable: false,
+                        evidence: "Pattern matched in error message but no execution evidence - tool properly validated input",
+                    };
+                }
             }
+            return {
+                isVulnerable: true,
+                evidence: `Tool executed malicious instruction: found evidence matching ${payload.evidence} - ${payload.description}`,
+            };
         }
-        return { detected: false, failureMode: "UNKNOWN" };
-    }
-    /**
-     * Check if response is from a creation/modification operation
-     */
-    isCreationResponse(responseText) {
-        const creationPatterns = [
-            /successfully\s+created/i,
-            /database\s+created/i,
-            /page\s+created/i,
-            /resource\s+created/i,
-            /\bcreate\s+table\b/i,
-            /\binsert\s+into\b/i,
-            /"id"\s*:\s*"[a-f0-9-]{36}"/i,
-            /"object"\s*:\s*"(page|database)"/i,
-            /collection:\/\//i,
-            /successfully\s+(added|inserted|updated|modified)/i,
-            /resource\s+id:\s*[a-f0-9-]/i,
-            /"created_time"/i,
-            /"last_edited_time"/i,
-        ];
-        return creationPatterns.some((pattern) => pattern.test(responseText));
-    }
-    /**
-     * Extract response content
-     */
-    extractResponseContent(response) {
-        if (response.content && Array.isArray(response.content)) {
-            return response.content
-                .map((c) => c.type === "text" ? c.text : "")
-                .join(" ");
-        }
-        return String(response.content || "");
+        // Fall back to injection response analysis
+        return this.analyzeInjectionResponse(response, payload.payload);
     }
     /**
-     * Extract error info from response
+     * Analyze injection response (fallback logic)
      */
-    extractErrorInfo(response) {
-        const content = this.extractResponseContent(response);
-        try {
-            const parsed = JSON.parse(content);
-            if (parsed.error) {
-                return {
-                    code: parsed.error.code || parsed.code,
-                    message: parsed.error.message || parsed.message,
-                };
-            }
-            return { code: parsed.code, message: parsed.message };
-        }
-        catch {
-            // Check for MCP error format in text
-            const mcpMatch = content.match(/MCP error (-?\d+):\s*(.*)/i);
-            if (mcpMatch) {
-                return { code: parseInt(mcpMatch[1]), message: mcpMatch[2] };
-            }
-            return {};
+    analyzeInjectionResponse(response, _payload) {
+        const analysis = this.executionDetector.analyzeInjectionResponse(this.extractResponseContent(response), (text) => this.safeDetector.isReflectionResponse(text));
+        if (analysis.isVulnerable) {
+            return {
+                isVulnerable: true,
+                evidence: analysis.evidence ||
+                    "Tool executed instruction: found execution keywords",
+            };
         }
+        return { isVulnerable: false };
     }
 }