@privacyscrubber/sdk 2.4.7 → 2.4.9

package/index.js

@@ -41,7 +41,7 @@ export const SDK_VERSION = (() => {
             return pkgJson.version;
         }
     } catch (_) {}
-    return '2.4.7';
+    return '2.4.9';
 })();
 
 /**
@@ -134,7 +134,7 @@ export function validateLicenseKey(key) {
  * @param {Record<string, string>|string[]} currentTokenMap 
  * @returns {object}
  */
-export function buildAuditTelemetry(currentTokenMap = {}) {
+export function buildAuditTelemetry(currentTokenMap = {}, extraStats = {}) {
     const entities = {};
     let totalCount = 0;
 
@@ -144,8 +144,9 @@ export function buildAuditTelemetry(currentTokenMap = {}) {
             if (!t) continue;
             const tokenStr = typeof t === 'string' ? t : ((t && (t.token || t.mask)) ? String(t.token || t.mask) : '');
             if (!tokenStr) continue;
-            const match = tokenStr.match(/\[([A-Z0-9_]+)_\d+\]/);
-            const baseType = match ? match[1] : (tokenStr.replace(/\[|\]/g, '').replace(/_[0-9]+$/, '') || 'CUSTOM');
+            const match = tokenStr.match(/\[([A-Za-z0-9_]+)_\d+\]/);
+            const rawBase = match ? match[1] : tokenStr.replace(/\[|\]|<|>|\{|\}|__/g, '').replace(/_[0-9]+$/, '');
+            const baseType = (rawBase.replace(/[^A-Za-z0-9_]/g, '') || 'CUSTOM').toUpperCase();
             entities[baseType] = (entities[baseType] || 0) + 1;
             totalCount++;
         }
@@ -155,7 +156,7 @@ export function buildAuditTelemetry(currentTokenMap = {}) {
     let riskLevel = 'LOW EXPOSURE';
 
     if (totalCount > 0) {
-        const hasHighRisk = types.some(t => ['ID', 'SSN', 'CREDIT_CARD', 'PASSPORT', 'BANK', 'API_KEY', 'SECRET', 'PASSWORD', 'MRN', 'KEY', 'AWS_KEY', 'STRIPE_KEY', 'JWT_TOKEN', 'API_TOKEN'].includes(t));
+        const hasHighRisk = types.some(t => ['ID', 'SSN', 'CREDIT_CARD', 'PASSPORT', 'BANK', 'API_KEY', 'SECRET', 'PASSWORD', 'MRN', 'KEY', 'PHI', 'PRIVILEGE', 'AWS_KEY', 'STRIPE_KEY', 'JWT_TOKEN', 'API_TOKEN'].includes(t));
         if (hasHighRisk || types.length >= 3 || totalCount >= 10) {
             riskLevel = 'CRITICAL (HIGH EXPOSURE)';
         } else if (types.length >= 2 || totalCount >= 3) {
@@ -180,10 +181,10 @@ export function buildAuditTelemetry(currentTokenMap = {}) {
         frameworksSet.add('PCI DSS v4.0');
         frameworksSet.add('IRC § 7216 & GLBA');
     }
-    if (types.some(t => ['MRN', 'HEALTH', 'MEDICAL', 'PATIENT', 'DIAGNOSIS', 'NPI', 'DEA'].includes(t))) {
+    if (types.some(t => ['PHI', 'MRN', 'HEALTH', 'MEDICAL', 'PATIENT', 'DIAGNOSIS', 'NPI', 'DEA'].includes(t))) {
         frameworksSet.add('HIPAA §164.514');
     }
-    if (types.some(t => ['CASE_NUMBER', 'DOCKET', 'PRIVILEGED', 'LEGAL', 'ATTORNEY', 'COURT', 'CONTRACT', 'NDA'].includes(t))) {
+    if (types.some(t => ['PRIVILEGE', 'CASE_NUMBER', 'DOCKET', 'PRIVILEGED', 'LEGAL', 'ATTORNEY', 'COURT', 'CONTRACT', 'NDA'].includes(t))) {
         frameworksSet.add('FRE 502 (Privilege)');
     }
     if (types.some(t => ['API_KEY', 'SECRET', 'PASSWORD', 'TOKEN', 'KEY', 'AWS_KEY', 'JWT_TOKEN', 'API_TOKEN', 'CREDENTIAL'].includes(t))) {
@@ -193,6 +194,26 @@ export function buildAuditTelemetry(currentTokenMap = {}) {
         frameworksSet.add('EU AI Act (Art. 50)');
     }
 
+    const frameworksList = Array.from(frameworksSet);
+    const getFrameworkSlug = (name) => {
+        const lower = name.toLowerCase();
+        if (lower.includes('gdpr')) return 'gdpr';
+        if (lower.includes('ccpa')) return 'ccpa';
+        if (lower.includes('soc')) return 'soc2';
+        if (lower.includes('iso')) return 'iso27001';
+        if (lower.includes('pci')) return 'pci';
+        if (lower.includes('hipaa')) return 'hipaa';
+        if (lower.includes('nist')) return 'nist';
+        if (lower.includes('irc') || lower.includes('7216')) return 'irc7216';
+        if (lower.includes('fre') || lower.includes('502')) return 'fre502';
+        if (lower.includes('ai act')) return 'euaiact';
+        return lower.replace(/[^a-z0-9]/g, '');
+    };
+    const triggeredFrameworks = frameworksList.map(name => ({
+        name,
+        id: getFrameworkSlug(name)
+    }));
+
     const ipHoldings = {
         standard: 'IETF draft-sibiryakov-ztds-protocol-01',
         standardUrl: 'https://datatracker.ietf.org/doc/draft-sibiryakov-ztds-protocol/',
@@ -244,12 +265,33 @@ export function buildAuditTelemetry(currentTokenMap = {}) {
         }
     ];
 
+    const rawLicenseTier = String(extraStats?.licenseTier || '').trim().toUpperCase();
+    const isOrganizationalScope = ['TEAMS', 'SDK', 'OEM', 'ENTERPRISE', 'ENTERPRISE_AIRGAP'].includes(rawLicenseTier);
+    const auditScope = isOrganizationalScope ? {
+        tier: rawLicenseTier,
+        isOrganizational: true,
+        badge: 'ORGANIZATIONAL MULTI-SEAT COMPLIANCE SCOPE',
+        coverage: 'Covers department, clinic, or firm personnel under unified ZTDS policy governance.',
+        upgradeRecommended: false,
+        upgradeUrl: null
+    } : {
+        tier: rawLicenseTier || 'SINGLE_OPERATOR',
+        isOrganizational: false,
+        badge: 'SINGLE-OPERATOR PERSONAL SCOPE',
+        coverage: 'Valid for individual workstation only. Multi-employee clinics, CPA firms, and departments require TEAMS ($99/mo flat, unlimited seats) for organizational audit coverage.',
+        upgradeRecommended: true,
+        upgradeUrl: 'https://privacyscrubber.com/pricing/?tier=teams&utm_source=audit_receipt&utm_medium=scope_upsell'
+    };
+
     return {
         totalCount,
+        totalDetected: totalCount,
         entities,
         types,
         riskLevel,
-        frameworksList: Array.from(frameworksSet),
+        frameworksList,
+        triggeredFrameworks,
+        auditScope,
         statutoryCitations,
         ipHoldings,
         agenticProtocol,
@@ -303,6 +345,19 @@ export function _resetCommunityNoticeForTesting() {
     communityNoticeLogged = false;
 }
 
+export const FREE_PROD_SPECIALIZED_CALL_LIMIT = 250;
+let _freeProdSpecializedCalls = 0;
+let _prodProfileFallbackNoticeLogged = false;
+
+/**
+ * Resets the production specialized call quota state for testing purposes.
+ * @private
+ */
+export function _resetProdQuotaForTesting() {
+    _freeProdSpecializedCalls = 0;
+    _prodProfileFallbackNoticeLogged = false;
+}
+
 let devSecretNoticeLogged = false;
 
 function emitDevSecretNoticeIfNeeded(isPro, tokenMap = {}, limit = 15000) {
@@ -483,6 +538,82 @@ We request approval to acquire a **Developer SDK License** (\$299/month flat or
  *   auditReceipt: string
  * }}
  */
+const SDK_VERTICAL_DETECTORS = [
+    { profile: 'medical', label: 'Medical PHI (HIPAA)', regex: /\b(?:Patient\s*(?:Name|ID|DOB|Record)|MRN\b|Medical\s+Record|Attending\s+Physician|Chief\s+Complaint|Diagnosis\s*:|ICD-10|HIPAA|Clinical\s+Intake|Discharge\s+Summary|MedPod|NPI\s*(?:#|No|:)?\s*\d{10}|Medicare\s*(?:ID|Beneficiary|Claim)|Medicaid\s*ID|Prescription|Rx\s*#|Blood\s+Type|Vital\s+Signs|Radiology|Lab\s+Results|EHR\b|EMR\b)\b/gi },
+    { profile: 'accounting', label: 'Tax & Accounting', regex: /\b(?:Form\s*1040|Form\s*W-2|Form\s*1099|Adjusted\s+Gross\s+Income|Assessable\s+Income|Employer\s+Identification\s+Number|Schedule\s+C\b|Steuer-ID|USt-IdNr|Tax\s+Return|General\s+Ledger|Withholding\s+Allowance|Total\s+Tax\s+Liability|ABN\b|TFN\b|Tax\s+File\s+Number|Australian\s+Business\s+Number|Business\s+Activity\s+Statement|PAYG\s+(?:Withholding|Instalment)|Franking\s+Credits?)\b/gi },
+    { profile: 'wealthmgmt', label: 'Wealth Management', regex: /\b(?:Revocable\s+Living\s+Trust|Irrevocable\s+(?:Generation-Skipping\s+)?Trust|Grantor\s*:|Settlor\s*:|Trustee\s*:|Beneficiary\s+Designation|Roth\s+IRA|Asset\s+Allocation|FBO\b|CUSIP\s*:|Grantor\s+Retained\s+Annuity\s+Trust|AUM\s*:)\b/gi },
+    { profile: 'legal', label: 'Legal & Contracts', regex: /\b(?:Non-Disclosure\s+Agreement|Attorney-Client\s+Privilege|Docket\s+No|Hereinafter\s+referred|Indemnification\s+Clause|Governing\s+Law\s+and\s+Jurisdiction|Plaintiff\b|Defendant\b|Subpoena\b|Deposition\b|State\s+Bar\s*#)\b/gi },
+    { profile: 'finance', label: 'Finance & Banking', regex: /\b(?:Wire\s+Transfer\s+Authorization|SWIFT\s*\/?\s*BIC\s*:|IBAN\s*:\s*[A-Z]{2}\d{2}|ABA\s+Routing|ACH\s+Transfer|PCI-DSS|Primary\s+Account\s+Number)\b/gi },
+    { profile: 'hr', label: 'HR & Resumes', regex: /\b(?:Employee\s*ID|Performance\s+Improvement\s+Plan|Payroll\s+Register|Gross\s+Pay\b|Net\s+Pay\b|Candidate\s+Evaluation|Compensation\s+Band|Sozialversicherungsnummer|Personalnummer|EEID\s*:)\b/gi },
+    { profile: 'realestate', label: 'Real Estate', regex: /\b(?:Assessor\s+Parcel\s+Number|APN\s*:|MLS\s*#|Closing\s+Disclosure|Escrow\s+(?:Account|Number|Officer)|Property\s+Deed|Title\s+Insurance|Lease\s+Application)\b/gi },
+    { profile: 'dev', label: 'API & Dev Logs', regex: /\b(?:BEGIN\s+(?:RSA|OPENSSH|EC|PRIVATE)\s+PRIVATE\s+KEY|AWS_SECRET_ACCESS_KEY|DATABASE_URL\s*=|GITHUB_TOKEN\s*=|Bearer\s+eyJ[A-Za-z0-9_-]+)\b/g },
+    { profile: 'sales', label: 'Sales & CRM', regex: /\b(?:CRM\s+(?:Opportunity|Export)|HubSpot\b|Salesforce\b|MEDDPICC\b|Primary\s+Buyer\s*:|MSA\s+Redlines|Deal\s+Stage|Pipeline\s+Forecast|Opportunity\s+ID|Account\s+Executive)\b/gi },
+    { profile: 'engineering', label: 'DevOps & Infra', regex: /\b(?:Incident\s+Postmortem|Kubeconfig|Kubernetes\s+Cluster|Helm\s+Chart|Terraform\s+State|DMAIC\b|Bill\s+of\s+Materials|BOM\s+Spec|CNC\s+Machining|Shop\s+Floor\s+Batch)\b/gi },
+    { profile: 'insurance', label: 'Insurance & Claims', regex: /\b(?:Claim\s+Settlement|Policyholder\s*:|Policy\s*#?POL-|Claim\s+Number|Adjuster\s*:|Loss\s+Location|Loss\s+Run|Subrogation\b|FNOL\b|Deductible\s*:)\b/gi },
+    { profile: 'pharma', label: 'Pharma & Trials', regex: /\b(?:Clinical\s+Trial(?:\s+Protocol)?|NCT\d{8}\b|IRB\s+Protocol|FDA\s+1572|Adverse\s+Event|Pharmacovigilance|Principal\s+Investigator|Cohort\s+[A-Z]\b|Blinded\s+Study|CDISC\b|IND\s*(?:#|Protocol))\b/gi },
+    { profile: 'underwriting', label: 'Loan Underwriting & Income', regex: /\b(?:Loan\s+Underwriting|Mortgage\s+Appraisal|Loan\s+Application\s+#?LOS|FICO\s*:\s*\d{3}|Appraised\s+Value\s*:|80%\s+LTV|Monthly\s+Qualifying\s*:|YTD\s+Earnings|Borrower\s*:)\b/gi },
+    { profile: 'security', label: 'Security & SIEM', regex: /\b(?:SIEM\s+Alert|CVE-\d{4}-\d{4,7}\b|Threat\s*:\s*CRITICAL|Unauthorized\s+S3\s+bucket|SOC\s+Incident|Firewall\s+Rule\s+Violation|Intrusion\s+Detection)\b/gi },
+    { profile: 'support', label: 'Support & Tickets', regex: /\b(?:Support\s+Ticket|Ticket\s*#INC-|Zendesk\b|Intercom\b|Billing\s+Discrepancy|SLA\s+Escalation|CSAT\s+Score|Customer\s+Support\s+Escalation)\b/gi },
+    { profile: 'compliance', label: 'Compliance Audit', regex: /\b(?:Suspicious\s+Activity\s+Report|FinCEN\s+SAR|BSA\s+Officer|31\s+CFR\b|ROPA\s+Register|Data\s+Protection\s+Impact\s+Assessment|SOC\s*2\s+Type\s+II|ISO\s*27001\s+Audit|CMMC\s+Level|ITAR\s+Controlled)\b/gi },
+    { profile: 'ccpa', label: 'CCPA / CPRA Privacy', regex: /\b(?:CCPA\s*\/?\s*CPRA|1798\.105|Right\s+to\s+Delete|Opt-out\s+of\s+Sale|Ticket\s*#?DSAR-)\b/gi },
+    { profile: 'bizops', label: 'Biz Ops & M&A', regex: /\b(?:Vendor\s+Master\s+Agreement|Bill\s+of\s+Lading|Customs\s+Declaration|Purchase\s+Order\s*#?PO-|Supply\s+Chain\s+Manifest)\b/gi },
+    { profile: 'biotech', label: 'Biotech & Genomics', regex: /\b(?:Sequencing\s+Platform|Illumina\s+NovaSeq|Flowcell\s*:|FastQ\s+Sample|ClinVar\b|HGVS\s+Variant|Genomic\s+Sequencing|Bioinformatics\s+Intake|BRCA[12]\s+c\.)\b/gi },
+    { profile: 'automotive', label: 'Automotive & EV', regex: /\b(?:Connected\s+Vehicle\s+Telematics|ECU\s+Serial|Telematics\s+Device|ASIL-D\b|DTC\s+P\d{4}\b|CAN-Bus\s+Log)\b/gi },
+    { profile: 'energy', label: 'Energy & Smart Grid', regex: /\b(?:Critical\s+Infrastructure\s+SCADA|Substation\s+ID|Feeder\s+ID|Smart\s+Meter\s*:|ESI\s+ID\s*:|NERC\s+CIP|BES\s+Cyber\s+Asset)\b/gi },
+    { profile: 'hospitality', label: 'Hospitality & Travel', regex: /\b(?:Airline\s+Passenger\s+Dossier|PNR\s*:\s*[A-Z0-9]{6}|e-Ticket\s*:|Boarding\s+Pass\s*:|Known\s+Traveler\s+KTN|Guest\s+Folio|Marriott\s+Bonvoy|SkyMiles)\b/gi },
+    { profile: 'telecom', label: 'Telecom & Call Center', regex: /\b(?:Telecom\s+Subscriber\s+CDR|MSISDN\s*:|IMSI\s*:\s*\d{14,15}|IMEI\s*:\s*\d{14,16}|Cell\s+Tower\s+ID|SIP\s+Signaling)\b/gi },
+    { profile: 'academic', label: 'Academic & Research', regex: /\b(?:FERPA\s+Student\s+Record|Student\s+ID\s*:\s*#?STU-|Disciplinary\s+Record|IEP\s+Evaluation|IRB\s+Approval)\b/gi },
+    { profile: 'agents', label: 'AI Agents & Prompts', regex: /\b(?:Autonomous\s+Financial\s+Procurement\s+Agent|AgentID\s*:\s*agt_|Encrypted\s+Tool\s+Token|Agent\s+Memory\s*>)\b/gi }
+];
+
+/**
+ * Automatically detects the most relevant industry compliance profile from text content.
+ * Executes in <0.05ms in local RAM with zero network egress.
+ *
+ * @param {string} text - Input document or prompt text.
+ * @returns {{
+ *   detectedProfile: string,
+ *   label: string,
+ *   confidence: 'high'|'medium'|'low',
+ *   score: number,
+ *   alternatives: Array<{ profile: string, label: string, score: number }>
+ * }}
+ */
+export function detectProfile(text) {
+    if (!text || typeof text !== 'string') {
+        return { detectedProfile: 'general', label: 'General PII', confidence: 'low', score: 0, alternatives: [] };
+    }
+    const sampleSlice = text.length > 16384 ? text.slice(0, 16384) : text;
+    const matches = [];
+    for (let i = 0; i < SDK_VERTICAL_DETECTORS.length; i++) {
+        const c = SDK_VERTICAL_DETECTORS[i];
+        c.regex.lastIndex = 0;
+        const found = sampleSlice.match(c.regex);
+        if (found && found.length > 0) {
+            matches.push({
+                profile: c.profile,
+                label: c.label,
+                score: found.length,
+                priority: i
+            });
+        }
+    }
+    if (matches.length === 0) {
+        return { detectedProfile: 'general', label: 'General PII', confidence: 'low', score: 0, alternatives: [] };
+    }
+    matches.sort((a, b) => (b.score !== a.score) ? (b.score - a.score) : (a.priority - b.priority));
+    const primary = matches[0];
+    const alternatives = matches.slice(1, 4).map(({ profile, label, score }) => ({ profile, label, score }));
+    const confidence = primary.score >= 3 ? 'high' : (primary.score >= 2 ? 'medium' : 'low');
+    return {
+        detectedProfile: primary.profile,
+        label: primary.label,
+        confidence,
+        score: primary.score,
+        alternatives
+    };
+}
+
 export function sanitize(text, options = {}) {
     if (!text || typeof text !== 'string') {
         return {
@@ -501,7 +632,7 @@ export function sanitize(text, options = {}) {
     }
 
     const {
-        profile = 'General',
+        profile: rawProfileOption = 'General',
         licenseKey = (typeof process !== 'undefined' && (process.env?.PRIVACYSCRUBBER_KEY || process.env?.PRIVACYSCRUBBER_LICENSE_KEY)) || '',
         customRules = [],
         tokenLabelMap = {},
@@ -510,9 +641,17 @@ export function sanitize(text, options = {}) {
         throwOnLimit = true,
         onLimit = 'throw',
         detectSecrets = false,
-        ignoreList = null
+        ignoreList = null,
+        entities = options.entityTypes || options.enabledEntities || null,
+        highRecallNames = Boolean(options.highRecallNames || options.nameRecall === 'high')
     } = options;
 
+    const isAutoProfile = Boolean(
+        (typeof rawProfileOption === 'string' && rawProfileOption.trim().toLowerCase() === 'auto') ||
+        options.autoDetectProfile === true
+    );
+    const autoDetected = isAutoProfile ? detectProfile(text) : null;
+
     const licenseCheck = validateLicense(licenseKey);
     const rawTier = licenseCheck.valid ? (licenseCheck.tier || 'PRO') : 'FREE';
     const isExpiredKey = !licenseCheck.valid && typeof licenseCheck.error === 'string' && /expired/i.test(licenseCheck.error);
@@ -534,6 +673,21 @@ export function sanitize(text, options = {}) {
     const isSdkTier = licenseCheck.valid && ['SDK', 'OEM', 'ENTERPRISE'].includes(rawTier);
     const isSeatTier = licenseCheck.valid && ['PRO', 'TEAMS'].includes(rawTier);
 
+    const isPro = isSdkTier;
+    const tier = isPro ? rawTier : (isSeatTier ? rawTier : 'FREE');
+
+    let autoSuggestedOnly = false;
+    let profile = rawProfileOption;
+    if (isAutoProfile && autoDetected) {
+        if (!isPro && text.length > 5000 && text.length <= 15000 && autoDetected.detectedProfile !== 'general') {
+            // Non-blocking PLG fallback: keep General for Free 5k-15k texts so auto-mode never breaks free calls
+            profile = 'General';
+            autoSuggestedOnly = true;
+        } else {
+            profile = autoDetected.detectedProfile;
+        }
+    }
+
     if (isSeatTier) {
         const keyPreview = licenseKey ? `${licenseKey.slice(0, 12)}...` : '';
         const errorMsg = `[PrivacyScrubber] License tier mismatch: ${rawTier} license (${keyPreview}) is licensed strictly for human seats (Web App & Chrome Extension). Backend automation, microservices, and CRM pipelines require a Developer SDK license ($299/mo). Upgrade: https://privacyscrubber.com/pricing?tier=sdk&src=seat_tier_mismatch`;
@@ -547,9 +701,6 @@ export function sanitize(text, options = {}) {
         }
     }
 
-    const isPro = isSdkTier;
-    const tier = isPro ? rawTier : (isSeatTier ? rawTier : 'FREE');
-
     emitCommunityNoticeIfNeeded(isPro);
 
     const isSpecialized = profile && profile.toLowerCase() !== 'general';
@@ -583,8 +734,48 @@ export function sanitize(text, options = {}) {
     });
 
     // Auto-inject DevOps Secrets rules when in Dev profile or when detectSecrets is requested
-    const activeRules = [...safeCustomRules];
-    if (profile.toLowerCase() === 'dev' || detectSecrets === true) {
+    const isProdEnv = typeof process !== 'undefined' && process.env?.NODE_ENV === 'production';
+    const hasCustomEnterpriseConfig = isSpecialized || safeCustomRules.length > 0 || (Array.isArray(ignoreList) && ignoreList.length > 0);
+    let profileFallback = false;
+    let effectiveProfile = profile;
+    let effectiveCustomRules = safeCustomRules;
+    let effectiveIgnoreList = ignoreList;
+
+    if (!isPro && isProdEnv && hasCustomEnterpriseConfig) {
+        _freeProdSpecializedCalls++;
+        if (_freeProdSpecializedCalls > FREE_PROD_SPECIALIZED_CALL_LIMIT) {
+            profileFallback = true;
+            effectiveProfile = 'General';
+            effectiveCustomRules = [];
+            effectiveIgnoreList = null;
+            if (!_prodProfileFallbackNoticeLogged && typeof process !== 'undefined' && process.stderr && typeof process.stderr.write === 'function') {
+                _prodProfileFallbackNoticeLogged = true;
+                process.stderr.write(
+                    `\n[@privacyscrubber/sdk] Free Production Quota (${FREE_PROD_SPECIALIZED_CALL_LIMIT} specialized calls) reached for profile '${profile}'.\n` +
+                    `[@privacyscrubber/sdk] Gracefully falling back to 'General' profile (Core PII remains protected, 0 downtime).\n` +
+                    `[@privacyscrubber/sdk] Unlock unlimited specialized profiles, custom rules & ignoreLists in production ($299/mo): https://privacyscrubber.com/pricing?tier=sdk&src=prod_profile_fallback (7-Day Trial: npx @privacyscrubber/sdk trial)\n\n`
+                );
+            }
+        }
+    }
+
+    let normalizedEntities = null;
+    if (entities) {
+        const rawArr = entities instanceof Set ? Array.from(entities) : (Array.isArray(entities) ? entities : [entities]);
+        normalizedEntities = rawArr.filter(Boolean).map(e => {
+            const u = String(e).trim().toUpperCase();
+            if (u === 'PERSON' || u === 'PERSONS' || u === 'PEOPLE' || u === 'NAMES') return 'NAME';
+            if (u === 'EMAILS') return 'EMAIL';
+            if (u === 'PHONES' || u === 'TELEPHONE') return 'PHONE';
+            if (u === 'IDS' || u === 'IDENTIFIER' || u === 'IDENTIFIERS') return 'ID';
+            if (u === 'ADDRESSES' || u === 'LOCATION' || u === 'LOCATIONS') return 'ADDRESS';
+            return u;
+        });
+        if (normalizedEntities.length === 0) normalizedEntities = null;
+    }
+
+    const activeRules = [...effectiveCustomRules];
+    if ((effectiveProfile.toLowerCase() === 'dev' || detectSecrets === true) && (!normalizedEntities || normalizedEntities.includes('SECRET') || normalizedEntities.includes('CUSTOM'))) {
         DEVOPS_SECRETS_DETECTOR.forEach(rule => {
             activeRules.push({
                 name: rule.name,
@@ -605,10 +796,12 @@ export function sanitize(text, options = {}) {
         textToScrub,
         activeRules,
         tokenLabelMap,
-        profile,
+        effectiveProfile,
         existingSessionMap,
         isPro,
-        ignoreList
+        effectiveIgnoreList,
+        normalizedEntities,
+        Boolean(highRecallNames)
     );
 
     const endTime = (typeof performance !== 'undefined' && typeof performance.now === 'function') 
@@ -616,7 +809,7 @@ export function sanitize(text, options = {}) {
         : Date.now();
 
     const executionMs = Math.round((endTime - startTime) * 100) / 100;
-    const telemetry = buildAuditTelemetry(rawResult.tokenMap || {});
+    const telemetry = buildAuditTelemetry(rawResult.tokenMap || {}, { licenseTier: tier });
     const auditReceipt = generateReceipt ? formatAuditReceipt(telemetry) : '';
 
     emitDevSecretNoticeIfNeeded(isPro, rawResult.tokenMap, limit);
@@ -644,6 +837,10 @@ export function sanitize(text, options = {}) {
             limit: limit,
             wasTruncated: wasTruncated,
             truncated: wasTruncated,
+            profileFallback: profileFallback,
+            effectiveProfile: effectiveProfile,
+            specializedCallsUsed: _freeProdSpecializedCalls,
+            specializedCallsLimit: isPro ? Infinity : FREE_PROD_SPECIALIZED_CALL_LIMIT,
             ...(isSeatTier && {
                 upgradeUrl: `https://privacyscrubber.com/pricing?tier=sdk&profile=${encodeURIComponent(String(profile || 'General').toLowerCase())}&src=seat_tier_mismatch`,
                 trialUrl: `https://privacyscrubber.com/pricing?tier=sdk&trial=true&profile=${encodeURIComponent(String(profile || 'General').toLowerCase())}&src=seat_tier_mismatch`,
@@ -651,12 +848,20 @@ export function sanitize(text, options = {}) {
                 recommendation: `Your ${rawTier} license is valid for interactive browser seats only. Upgrade to Developer SDK ($299/mo) for unlimited backend server throughput: https://privacyscrubber.com/pricing?tier=sdk`
             }),
             ...(!isPro && !isSeatTier && {
-                upgradeUrl: `https://privacyscrubber.com/pricing?tier=sdk&profile=${encodeURIComponent(String(profile || 'General').toLowerCase())}&src=dev_upsell`,
-                trialUrl: `https://privacyscrubber.com/pricing?tier=sdk&trial=true&profile=${encodeURIComponent(String(profile || 'General').toLowerCase())}&src=dev_upsell`,
+                upgradeUrl: `https://privacyscrubber.com/pricing?tier=sdk&profile=${encodeURIComponent(String(profile || 'General').toLowerCase())}&src=${profileFallback ? 'prod_profile_fallback' : 'dev_upsell'}`,
+                trialUrl: `https://privacyscrubber.com/pricing?tier=sdk&trial=true&profile=${encodeURIComponent(String(profile || 'General').toLowerCase())}&src=${profileFallback ? 'prod_profile_fallback' : 'dev_upsell'}`,
                 trialCommand: 'npx @privacyscrubber/sdk trial',
                 recommendation: 'Upgrade to Developer SDK ($299/mo) for unlimited backend text processing, custom regex rules, and specialized compliance profiles.'
             })
         },
+        profileDetection: {
+            autoDetected: isAutoProfile,
+            detectedProfile: autoDetected ? autoDetected.detectedProfile : String(effectiveProfile || 'General').toLowerCase(),
+            label: autoDetected ? autoDetected.label : effectiveProfile,
+            confidence: autoDetected ? autoDetected.confidence : 'high',
+            suggestedOnly: autoSuggestedOnly,
+            alternatives: autoDetected ? autoDetected.alternatives : []
+        },
         telemetry: telemetry,
         auditReceipt: auditReceipt
     };
@@ -924,6 +1129,8 @@ export class PrivacyScrubberEngine {
         this.generateReceipt = config.generateReceipt !== false;
         this.sessionMap = { ...(config.initialSessionMap || {}) };
         this.ignoreList = new Set(config.ignoreList || []);
+        this.entities = config.entities || config.entityTypes || config.enabledEntities || null;
+        this.highRecallNames = Boolean(config.highRecallNames || config.nameRecall === 'high');
         this.maxSessionTokens = config.maxSessionTokens || 10000;
         
         this.license = validateLicense(this.licenseKey);
@@ -965,7 +1172,9 @@ export class PrivacyScrubberEngine {
             generateReceipt: options.generateReceipt !== undefined ? options.generateReceipt : this.generateReceipt,
             throwOnLimit: options.throwOnLimit !== undefined ? options.throwOnLimit : true,
             detectSecrets: options.detectSecrets !== undefined ? options.detectSecrets : false,
-            ignoreList: mergedIgnoreList.size > 0 ? mergedIgnoreList : null
+            ignoreList: mergedIgnoreList.size > 0 ? mergedIgnoreList : null,
+            entities: options.entities || options.entityTypes || options.enabledEntities || this.entities,
+            highRecallNames: options.highRecallNames !== undefined ? Boolean(options.highRecallNames) : (options.nameRecall === 'high' ? true : this.highRecallNames)
         };
 
         const result = sanitize(text, mergedOptions);
@@ -2505,13 +2714,13 @@ export function buildStreamReplacer(tokenMap = {}, options = {}) {
 
     for (const k of sortedKeys) {
         const val = tokenMap[k];
-        const inner = k.replace(/^\[|<|\{\{?|__|\]|>|\}\}?|__/g, '');
+        const inner = k.replace(/^(?:\[|<|\{\{?|__|【|［|〔|「|『)|(?:\]|>|\}\}?|__|】|］|〕|」|』)$/g, '');
         const match = inner.match(/^([A-Za-z_0-9]+?)[-_ ]?(\d+)$/);
 
         if (match) {
             const label = escapeRegex(match[1]).replace(/[_-]/g, '\\\\?[_-]');
             const idx = match[2];
-            const patternStr = '(?:(?:\\[|\\\\u005[Bb]|%5[Bb]|&(?:amp;)?#0*91;|&(?:amp;)?#[xX]0*5[Bb];|&(?:amp;)?lbrack;|【|［|〔|「|『)\\s*[*`~]*' + label + '[-_\\\\\\s]*0*' + idx + '[*`~]*[.,;:!?]?\\s*(?:\\]|\\\\u005[Dd]|%5[Dd]|&(?:amp;)?#0*93;|&(?:amp;)?#[xX]0*5[Dd];|&(?:amp;)?rbrack;|】|］|〕|」|』)|<\\s*' + label + '[-_\\\\\\s]*0*' + idx + '[.,;:!?]?\\s*>|\\{\\{?\\s*' + label + '[-_\\\\\\s]*0*' + idx + '[.,;:!?]?\\s*\\}\\}?|__\\s*' + label + '[-_\\\\\\s]*0*' + idx + '\\s*__|\\b' + label + '[-_\\s]+0*' + idx + '\\b)';
+            const patternStr = '(?:(?:\\[|\\\\u005[Bb]|%5[Bb]|&(?:amp;)?#0*91;|&(?:amp;)?#[xX]0*5[Bb];|&(?:amp;)?(?:lbrack|lsqb);|【|［|〔|「|『)\\s*[*`~]*' + label + '[-_\\\\\\s]*0*' + idx + '[*`~]*[.,;:!?]?\\s*(?:\\]|\\\\u005[Dd]|%5[Dd]|&(?:amp;)?#0*93;|&(?:amp;)?#[xX]0*5[Dd];|&(?:amp;)?(?:rbrack|rsqb);|】|］|〕|」|』)|<\\s*' + label + '[-_\\\\\\s]*0*' + idx + '[.,;:!?]?\\s*>|\\{\\{?\\s*' + label + '[-_\\\\\\s]*0*' + idx + '[.,;:!?]?\\s*\\}\\}?|__\\s*' + label + '[-_\\\\\\s]*0*' + idx + '\\s*__|\\b' + label + '[-_\\s]+0*' + idx + '\\b)';
             patterns.push(patternStr);
             const normKey = (match[1] + '_' + parseInt(idx, 10)).toUpperCase();
             lookup.set(normKey, val);
@@ -2541,8 +2750,8 @@ export function buildStreamReplacer(tokenMap = {}, options = {}) {
                 restored = lookup.get(coreMatch.toUpperCase());
             } else {
                 const inner = coreMatch
-                    .replace(/^(?:\[|\\u005[Bb]|%5[Bb]|&(?:amp;)?#0*91;|&(?:amp;)?#[xX]0*5[Bb];|&(?:amp;)?lbrack;|【|［|〔|「|『|<|\{\{?|__)|(?:\]|\\u005[Dd]|%5[Dd]|&(?:amp;)?#0*93;|&(?:amp;)?#[xX]0*5[Dd];|&(?:amp;)?rbrack;|】|］|〕|」|』|>|\}\}?|__)$/gi, '')
-                    .replace(/^[*`~\s]+|[.,;:!?*`~\s]+$/g, '')
+                    .replace(/^(?:\[|\\u005[Bb]|%5[Bb]|&(?:amp;)?#0*91;|&(?:amp;)?#[xX]0*5[Bb];|&(?:amp;)?(?:lbrack|lsqb);|【|［|〔|「|『|<|\{\{?|__)|(?:\]|\\u005[Dd]|%5[Dd]|&(?:amp;)?#0*93;|&(?:amp;)?#[xX]0*5[Dd];|&(?:amp;)?(?:rbrack|rsqb);|】|］|〕|」|』|>|\}\}?|__)$/gi, '')
+                    .replace(/^[*`_~\s]+|[.,;:!?*`_~\s]+$/g, '')
                     .trim();
                 const m = inner.match(/^([A-Za-z_0-9]+?)[-_\\\s]*(\d+)$/);
                 if (m) {
@@ -2655,18 +2864,37 @@ export class StreamingDetokenizer {
         // Candidate unclosed token at tail of buffer (inspect tail window up to 160 chars)
         // Matches incomplete tokens starting with [, 【, ［, <, {, __, %5B, \u005b, or HTML entity &#91; / &lbrack;
         const tailLen = 160;
-        const tail = this.pendingBuffer.length > tailLen ? this.pendingBuffer.slice(-tailLen) : this.pendingBuffer;
-        const offset = this.pendingBuffer.length > tailLen ? this.pendingBuffer.length - tailLen : 0;
-        const m = tail.match(/(?:(?:\[|【|［|〔|「|『|<|\{\{?|__?|%5[Bb]|\\u005[Bb]|&(?:amp;)?(?:#0*91;|#[xX]0*5[Bb];|lbrack;))[A-Za-z0-9_\-\u0410-\u044F\u0401\u0451 *`~.\\]{0,128}(?:%5?|&(?:a(?:m(?:p(?:;)?)?)?)?(?:#(?:0*9?3?|[xX]0*5?[Dd]?)?|r(?:b(?:r(?:a(?:c(?:k)?)?)?)?)?)?)?|%[0-9A-Fa-f]?|\\(?:u(?:0(?:0(?:5)?)?)?)?|&(?:a(?:m(?:p(?:;)?)?)?)?(?:#(?:0*9?1?|[xX]0*5?)?|l(?:b(?:r(?:a(?:c(?:k)?)?)?)?)?)?)$/);
+        const tailRegex = /(?:(?:\[|【|［|〔|「|『|<|\{\{?|__|%5[Bb]|\\u005[Bb]|&(?:amp;)?(?:#0*91;|#[xX]0*5[Bb];|lbrack;))[A-Za-z0-9_\-\u0410-\u044F\u0401\u0451 *`~.\\]{0,128}(?:%5?|&(?:a(?:m(?:p(?:;)?)?)?)?(?:#(?:0*9?3?|[xX]0*5?[Dd]?)?|r(?:b(?:r(?:a(?:c(?:k)?)?)?)?)?)?)?|%[0-9A-Fa-f]?|\\(?:u(?:0(?:0(?:5)?)?)?)?|&(?:a(?:m(?:p(?:;)?)?)?)?(?:#(?:0*9?1?|[xX]0*5?)?|l(?:b(?:r(?:a(?:c(?:k)?)?)?)?)?)?|_)$/;
+        let tail = this.pendingBuffer.length > tailLen ? this.pendingBuffer.slice(-tailLen) : this.pendingBuffer;
+        let offset = this.pendingBuffer.length > tailLen ? this.pendingBuffer.length - tailLen : 0;
+        let m = tail.match(tailRegex);
 
         if (m) {
             const candidate = m[0];
             // Closed double-underscore (__NAME_1__) or JSON Unicode (\u005bNAME_1\u005d) tokens match character class and can be emitted immediately.
-            const isClosedUnderscoreOrUnicode = (candidate.length >= 4 && candidate.startsWith('__') && candidate.endsWith('__')) || /\\u005[Dd]/.test(candidate);
-            if (isClosedUnderscoreOrUnicode && this.replacer(candidate) !== candidate) {
-                const toEmit = this.pendingBuffer;
-                this.pendingBuffer = '';
-                return this.replacer(toEmit);
+            const uClose = candidate.startsWith('__') ? candidate.indexOf('__', 2) : -1;
+            const uniMatch = candidate.match(/\\u005[Dd]/);
+            const closeLen = uClose !== -1 ? uClose + 2 : (uniMatch ? uniMatch.index + 6 : -1);
+            if (closeLen !== -1) {
+                const closedToken = candidate.slice(0, closeLen);
+                if (this.replacer(closedToken) !== closedToken) {
+                    const splitAfterToken = offset + m.index + closeLen;
+                    const prefixEmitted = this.replacer(this.pendingBuffer.slice(0, splitAfterToken));
+                    this.pendingBuffer = this.pendingBuffer.slice(splitAfterToken);
+                    if (!this.pendingBuffer) return prefixEmitted;
+                    tail = this.pendingBuffer.length > tailLen ? this.pendingBuffer.slice(-tailLen) : this.pendingBuffer;
+                    offset = this.pendingBuffer.length > tailLen ? this.pendingBuffer.length - tailLen : 0;
+                    m = tail.match(tailRegex);
+                    if (!m) {
+                        const rest = this.replacer(this.pendingBuffer);
+                        this.pendingBuffer = '';
+                        return prefixEmitted + rest;
+                    }
+                    const sIdx = offset + m.index;
+                    const safe = this.pendingBuffer.slice(0, sIdx);
+                    this.pendingBuffer = this.pendingBuffer.slice(sIdx);
+                    return prefixEmitted + (safe ? this.replacer(safe) : '');
+                }
             }
             const splitIdx = offset + m.index;
             const safeToEmit = this.pendingBuffer.slice(0, splitIdx);
@@ -3888,6 +4116,7 @@ import {
 
 export default {
     SDK_VERSION,
+    detectProfile,
     sanitize,
     sanitizeObject,
     scrub,
@@ -3940,6 +4169,8 @@ export default {
     scrubText,
     unscrubText,
     DEVOPS_SECRETS_DETECTOR,
+    FREE_PROD_SPECIALIZED_CALL_LIMIT,
+    _resetProdQuotaForTesting,
     stripExifMetadata,
     redactImageBoundingBoxes,
     generateSvgRedactionOverlay

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "@privacyscrubber/sdk",
-  "version": "2.4.7",
+  "version": "2.4.9",
   "description": "In-process consumer privacy defense and pre-emptive PII interception SDK for Node.js, Next.js, and AI pipelines. Zero data loss, zero third parties, sub-millisecond in-memory execution.",
   "main": "index.cjs",
   "module": "index.js",