palaryn 0.1.0

package/dist/src/dlp/scanner.js

@@ -0,0 +1,444 @@
+"use strict";
+Object.defineProperty(exports, "__esModule", { value: true });
+exports.DLPScanner = void 0;
+const crypto_1 = require("crypto");
+const patterns_1 = require("./patterns");
+/** Maximum recursion depth to guard against circular or deeply nested structures. */
+const MAX_SCAN_DEPTH = 32;
+/**
+ * DLPScanner detects secrets and PII in tool call arguments and outputs,
+ * and applies configurable redaction strategies (mask, hash, drop, tokenize).
+ */
+class DLPScanner {
+    constructor(config) {
+        this.config = config;
+    }
+    /**
+     * Main scan function -- recursively scans an object for secrets and PII.
+     *
+     * @param data - The data to scan (object, array, string, or primitive).
+     * @param basePath - Dot-notation prefix for paths within the data structure.
+     * @returns A DLPReport describing all detections, suggested redactions, and overall severity.
+     */
+    scan(data, basePath = '') {
+        // If DLP is disabled, return an empty report immediately
+        if (!this.config.enabled) {
+            return {
+                detected: [],
+                redactions: [],
+                severity: 'low',
+            };
+        }
+        const detected = [];
+        const redactions = [];
+        this.scanObject(data, basePath, detected, redactions, 0);
+        // Deduplicate detection names while preserving order
+        const uniqueDetected = [...new Set(detected)];
+        const severity = this.calculateSeverity(detected);
+        return {
+            detected: uniqueDetected,
+            redactions,
+            severity,
+        };
+    }
+    /**
+     * Scan a single string value for secrets and PII patterns.
+     *
+     * Every regex with the /g flag has its lastIndex reset before and after
+     * testing to avoid state leaking between calls.
+     */
+    scanString(value, path) {
+        const detected = [];
+        const redactions = [];
+        if (this.config.secrets_detection) {
+            for (const pattern of patterns_1.SECRET_PATTERNS) {
+                // Reset before use to ensure a clean match
+                pattern.pattern.lastIndex = 0;
+                if (pattern.pattern.test(value)) {
+                    // Extract the first match for masked preview
+                    pattern.pattern.lastIndex = 0;
+                    const match = pattern.pattern.exec(value);
+                    detected.push(pattern.name);
+                    redactions.push({
+                        path,
+                        method: this.config.default_redaction_method,
+                        original_type: pattern.name,
+                        masked_preview: match ? DLPScanner.maskValue(match[0]) : undefined,
+                    });
+                }
+                // Reset after use so the next invocation starts fresh
+                pattern.pattern.lastIndex = 0;
+            }
+        }
+        if (this.config.pii_detection) {
+            for (const pattern of patterns_1.PII_PATTERNS) {
+                pattern.pattern.lastIndex = 0;
+                if (pattern.pattern.test(value)) {
+                    pattern.pattern.lastIndex = 0;
+                    const match = pattern.pattern.exec(value);
+                    detected.push(pattern.name);
+                    redactions.push({
+                        path,
+                        method: this.config.default_redaction_method,
+                        original_type: pattern.name,
+                        masked_preview: match ? DLPScanner.maskValue(match[0]) : undefined,
+                    });
+                }
+                pattern.pattern.lastIndex = 0;
+            }
+        }
+        return { detected, redactions };
+    }
+    /** Mask a matched value, keeping first 3 and last 3 chars visible. */
+    static maskValue(value) {
+        if (value.length <= 8)
+            return value.slice(0, 2) + '***' + value.slice(-2);
+        return value.slice(0, 3) + '***' + value.slice(-3);
+    }
+    /**
+     * Recursively walk an unknown value, scanning every string leaf.
+     *
+     * A depth limit prevents stack overflow on circular or extremely deep structures.
+     */
+    scanObject(obj, path, detected, redactions, depth) {
+        // Guard against circular references / deeply nested objects
+        if (depth > MAX_SCAN_DEPTH) {
+            return;
+        }
+        if (obj === null || obj === undefined) {
+            return;
+        }
+        if (typeof obj === 'string') {
+            const result = this.scanString(obj, path);
+            detected.push(...result.detected);
+            redactions.push(...result.redactions);
+            return;
+        }
+        if (Array.isArray(obj)) {
+            for (let index = 0; index < obj.length; index++) {
+                const childPath = path ? `${path}[${index}]` : `[${index}]`;
+                this.scanObject(obj[index], childPath, detected, redactions, depth + 1);
+            }
+            return;
+        }
+        if (typeof obj === 'object') {
+            for (const [key, value] of Object.entries(obj)) {
+                const childPath = path ? `${path}.${key}` : key;
+                this.scanObject(value, childPath, detected, redactions, depth + 1);
+            }
+        }
+        // Non-string primitives (number, boolean, bigint, symbol, function) are ignored.
+    }
+    /**
+     * Apply a single redaction method to a string value.
+     *
+     * - `mask`     -- Keep the first 4 characters, replace the rest with 'x' (capped at 20 x's).
+     * - `hash`     -- SHA-256 hash prefix (16 hex chars) with a HASH: prefix.
+     * - `drop`     -- Replace entirely with `[REDACTED]`.
+     * - `tokenize` -- Replace with a deterministic placeholder token derived from MD5.
+     */
+    static redact(value, method) {
+        switch (method) {
+            case 'mask':
+                if (value.length <= 4)
+                    return 'xxxx';
+                return value.slice(0, 4) + 'x'.repeat(Math.min(value.length - 4, 20));
+            case 'hash':
+                return 'HASH:' + (0, crypto_1.createHash)('sha256').update(value).digest('hex').slice(0, 16);
+            case 'drop':
+                return '[REDACTED]';
+            case 'tokenize':
+                return '{{REDACTED_' + (0, crypto_1.createHash)('sha256').update(value).digest('hex').slice(0, 16) + '}}';
+            default:
+                return '[REDACTED]';
+        }
+    }
+    /**
+     * Apply a list of redactions to a data structure.
+     *
+     * Returns a deep clone of `data` with every path referenced in `redactions`
+     * replaced by the redacted version of the string found at that path.
+     *
+     * Paths use lodash-style dot/bracket notation:
+     *   - `"args.headers.Authorization"` -> obj.args.headers.Authorization
+     *   - `"items[0].value"`             -> obj.items[0].value
+     *
+     * If a path cannot be resolved (e.g. the structure was mutated), the
+     * redaction is silently skipped.
+     */
+    applyRedactions(data, redactions) {
+        if (redactions.length === 0) {
+            return data;
+        }
+        // Deep clone to avoid mutating the original data
+        const cloned = DLPScanner.deepClone(data);
+        for (const redaction of redactions) {
+            this.applyRedactionAtPath(cloned, redaction.path, redaction.method);
+        }
+        return cloned;
+    }
+    /**
+     * Navigate to `path` inside `root` and replace the leaf string value
+     * with its redacted form.
+     */
+    applyRedactionAtPath(root, path, method) {
+        const segments = DLPScanner.parsePath(path);
+        if (segments.length === 0) {
+            return;
+        }
+        // Walk to the parent of the target
+        let current = root;
+        for (let i = 0; i < segments.length - 1; i++) {
+            current = DLPScanner.getChild(current, segments[i]);
+            if (current === null || current === undefined) {
+                return; // path not resolvable
+            }
+        }
+        const lastSegment = segments[segments.length - 1];
+        const leafValue = DLPScanner.getChild(current, lastSegment);
+        if (typeof leafValue !== 'string') {
+            return; // nothing to redact
+        }
+        // Find all matching patterns in the leaf string and redact them
+        const redacted = this.redactMatchesInString(leafValue, method);
+        DLPScanner.setChild(current, lastSegment, redacted);
+    }
+    /**
+     * Redact all pattern matches within a string, replacing each match occurrence
+     * with the appropriate redacted form. Non-matching portions of the string are
+     * preserved. If no matches are found, the entire string is redacted as a
+     * fallback (e.g. when the string itself is the sensitive value).
+     */
+    redactMatchesInString(value, method) {
+        // Collect all match ranges across all enabled pattern sets
+        const allPatterns = [];
+        if (this.config.secrets_detection) {
+            for (const p of patterns_1.SECRET_PATTERNS) {
+                allPatterns.push(p.pattern);
+            }
+        }
+        if (this.config.pii_detection) {
+            for (const p of patterns_1.PII_PATTERNS) {
+                allPatterns.push(p.pattern);
+            }
+        }
+        const matches = [];
+        for (const regex of allPatterns) {
+            regex.lastIndex = 0;
+            let m;
+            while ((m = regex.exec(value)) !== null) {
+                matches.push({ start: m.index, end: m.index + m[0].length, text: m[0] });
+                // Prevent infinite loops on zero-length matches
+                if (m[0].length === 0) {
+                    regex.lastIndex++;
+                }
+            }
+            regex.lastIndex = 0;
+        }
+        if (matches.length === 0) {
+            // No specific matches found; redact the entire string as a fallback
+            return DLPScanner.redact(value, method);
+        }
+        // Sort matches by start position, then by length descending (prefer longer matches)
+        matches.sort((a, b) => a.start - b.start || b.end - a.end);
+        // Merge overlapping ranges and build the redacted string
+        let result = '';
+        let cursor = 0;
+        for (const match of matches) {
+            // Skip matches that overlap with already-processed ranges
+            if (match.start < cursor) {
+                continue;
+            }
+            // Append the non-sensitive text before this match
+            result += value.slice(cursor, match.start);
+            // Append the redacted match
+            result += DLPScanner.redact(match.text, method);
+            cursor = match.end;
+        }
+        // Append any remaining text after the last match
+        result += value.slice(cursor);
+        return result;
+    }
+    /**
+     * Parse a lodash-style path string into an array of segments.
+     *
+     * Examples:
+     *   `"args.headers.Authorization"` -> `["args", "headers", "Authorization"]`
+     *   `"items[0].value"`             -> `["items", "0", "value"]`
+     *   `"[0]"`                        -> `["0"]`
+     */
+    static parsePath(path) {
+        if (!path)
+            return [];
+        const segments = [];
+        let current = '';
+        for (let i = 0; i < path.length; i++) {
+            const char = path[i];
+            if (char === '.') {
+                if (current) {
+                    segments.push(current);
+                    current = '';
+                }
+            }
+            else if (char === '[') {
+                if (current) {
+                    segments.push(current);
+                    current = '';
+                }
+                // Read until closing bracket
+                const closingIndex = path.indexOf(']', i + 1);
+                if (closingIndex === -1) {
+                    // Malformed path -- treat rest as a single segment
+                    current = path.slice(i);
+                    break;
+                }
+                segments.push(path.slice(i + 1, closingIndex));
+                i = closingIndex; // the loop will increment past ']'
+            }
+            else {
+                current += char;
+            }
+        }
+        if (current) {
+            segments.push(current);
+        }
+        return segments;
+    }
+    /**
+     * Get a child value from an object or array given a single path segment.
+     */
+    static getChild(obj, segment) {
+        if (obj === null || obj === undefined)
+            return undefined;
+        if (Array.isArray(obj)) {
+            const index = parseInt(segment, 10);
+            if (Number.isNaN(index))
+                return undefined;
+            return obj[index];
+        }
+        if (typeof obj === 'object') {
+            return obj[segment];
+        }
+        return undefined;
+    }
+    /**
+     * Set a child value on an object or array given a single path segment.
+     */
+    static setChild(obj, segment, value) {
+        if (obj === null || obj === undefined)
+            return;
+        if (Array.isArray(obj)) {
+            const index = parseInt(segment, 10);
+            if (!Number.isNaN(index) && index >= 0 && index < obj.length) {
+                obj[index] = value;
+            }
+            return;
+        }
+        if (typeof obj === 'object') {
+            obj[segment] = value;
+        }
+    }
+    /**
+     * Deep clone a value using structured clone semantics.
+     *
+     * Falls back to JSON round-trip for environments where structuredClone is
+     * unavailable. Handles primitives, plain objects, arrays, and null/undefined.
+     */
+    static deepClone(data) {
+        if (data === null || data === undefined)
+            return data;
+        if (typeof data !== 'object')
+            return data;
+        // structuredClone is available in Node.js >= 17
+        if (typeof structuredClone === 'function') {
+            try {
+                return structuredClone(data);
+            }
+            catch {
+                // Fall through to JSON-based clone if structuredClone fails
+                // (e.g. on functions or symbols in the data)
+            }
+        }
+        try {
+            return JSON.parse(JSON.stringify(data));
+        }
+        catch {
+            // If JSON serialization fails (circular refs), do a manual shallow-ish clone
+            return data;
+        }
+    }
+    /**
+     * Determine the highest severity among all detected pattern names.
+     *
+     * Checks both SECRET_PATTERNS and PII_PATTERNS for each detection name
+     * and returns the highest severity found. Defaults to 'low' when no
+     * detections are present or none match a known pattern.
+     */
+    calculateSeverity(detected) {
+        if (detected.length === 0)
+            return 'low';
+        const allPatterns = [...patterns_1.SECRET_PATTERNS, ...patterns_1.PII_PATTERNS];
+        // Build a lookup from pattern name to severity
+        const severityMap = new Map();
+        for (const p of allPatterns) {
+            severityMap.set(p.name, p.severity);
+        }
+        const severityRank = {
+            low: 0,
+            medium: 1,
+            high: 2,
+        };
+        let highest = 'low';
+        for (const name of detected) {
+            const sev = severityMap.get(name);
+            if (sev && severityRank[sev] > severityRank[highest]) {
+                highest = sev;
+            }
+            // Short-circuit: can't get higher than 'high'
+            if (highest === 'high')
+                break;
+        }
+        return highest;
+    }
+    /**
+     * Calculate the Shannon entropy of a string.
+     *
+     * High-entropy strings (random-looking) are more likely to be secrets,
+     * API keys, or tokens even if they don't match a known pattern.
+     *
+     * @returns Entropy in bits per character (0 to ~log2(charset_size)).
+     */
+    static calculateEntropy(str) {
+        if (str.length === 0)
+            return 0;
+        const freq = {};
+        for (const c of str) {
+            freq[c] = (freq[c] || 0) + 1;
+        }
+        let entropy = 0;
+        const len = str.length;
+        for (const count of Object.values(freq)) {
+            const p = count / len;
+            entropy -= p * Math.log2(p);
+        }
+        return entropy;
+    }
+    /**
+     * Check if a string looks like a high-entropy secret.
+     *
+     * Requires both a minimum length (to ignore short random-ish strings like
+     * UUIDs in non-secret contexts) and an entropy above the threshold.
+     *
+     * @param str - The string to check.
+     * @param threshold - Minimum entropy in bits per character. Default 4.5
+     *   (base64-encoded random bytes typically have entropy ~5.2).
+     * @returns True if the string exceeds both the length and entropy thresholds.
+     */
+    static isHighEntropy(str, threshold = 4.5) {
+        if (str.length < 16)
+            return false;
+        return DLPScanner.calculateEntropy(str) >= threshold;
+    }
+}
+exports.DLPScanner = DLPScanner;
+//# sourceMappingURL=scanner.js.map

package/dist/src/dlp/scanner.js.map

@@ -0,0 +1 @@
+{"version":3,"file":"scanner.js","sourceRoot":"","sources":["../../../src/dlp/scanner.ts"],"names":[],"mappings":";;;AAAA,mCAAoC;AAGpC,yCAA2D;AAE3D,qFAAqF;AACrF,MAAM,cAAc,GAAG,EAAE,CAAC;AAE1B;;;GAGG;AACH,MAAa,UAAU;IAGrB,YAAY,MAAiB;QAC3B,IAAI,CAAC,MAAM,GAAG,MAAM,CAAC;IACvB,CAAC;IAED;;;;;;OAMG;IACH,IAAI,CAAC,IAAa,EAAE,WAAmB,EAAE;QACvC,yDAAyD;QACzD,IAAI,CAAC,IAAI,CAAC,MAAM,CAAC,OAAO,EAAE,CAAC;YACzB,OAAO;gBACL,QAAQ,EAAE,EAAE;gBACZ,UAAU,EAAE,EAAE;gBACd,QAAQ,EAAE,KAAK;aAChB,CAAC;QACJ,CAAC;QAED,MAAM,QAAQ,GAAa,EAAE,CAAC;QAC9B,MAAM,UAAU,GAAmB,EAAE,CAAC;QAEtC,IAAI,CAAC,UAAU,CAAC,IAAI,EAAE,QAAQ,EAAE,QAAQ,EAAE,UAAU,EAAE,CAAC,CAAC,CAAC;QAEzD,qDAAqD;QACrD,MAAM,cAAc,GAAG,CAAC,GAAG,IAAI,GAAG,CAAC,QAAQ,CAAC,CAAC,CAAC;QAE9C,MAAM,QAAQ,GAAG,IAAI,CAAC,iBAAiB,CAAC,QAAQ,CAAC,CAAC;QAElD,OAAO;YACL,QAAQ,EAAE,cAAc;YACxB,UAAU;YACV,QAAQ;SACT,CAAC;IACJ,CAAC;IAED;;;;;OAKG;IACK,UAAU,CAAC,KAAa,EAAE,IAAY;QAC5C,MAAM,QAAQ,GAAa,EAAE,CAAC;QAC9B,MAAM,UAAU,GAAmB,EAAE,CAAC;QAEtC,IAAI,IAAI,CAAC,MAAM,CAAC,iBAAiB,EAAE,CAAC;YAClC,KAAK,MAAM,OAAO,IAAI,0BAAe,EAAE,CAAC;gBACtC,2CAA2C;gBAC3C,OAAO,CAAC,OAAO,CAAC,SAAS,GAAG,CAAC,CAAC;gBAC9B,IAAI,OAAO,CAAC,OAAO,CAAC,IAAI,CAAC,KAAK,CAAC,EAAE,CAAC;oBAChC,6CAA6C;oBAC7C,OAAO,CAAC,OAAO,CAAC,SAAS,GAAG,CAAC,CAAC;oBAC9B,MAAM,KAAK,GAAG,OAAO,CAAC,OAAO,CAAC,IAAI,CAAC,KAAK,CAAC,CAAC;oBAC1C,QAAQ,CAAC,IAAI,CAAC,OAAO,CAAC,IAAI,CAAC,CAAC;oBAC5B,UAAU,CAAC,IAAI,CAAC;wBACd,IAAI;wBACJ,MAAM,EAAE,IAAI,CAAC,MAAM,CAAC,wBAAwB;wBAC5C,aAAa,EAAE,OAAO,CAAC,IAAI;wBAC3B,cAAc,EAAE,KAAK,CAAC,CAAC,CAAC,UAAU,CAAC,SAAS,CAAC,KAAK,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,SAAS;qBACnE,CAAC,CAAC;gBACL,CAAC;gBACD,sDAAsD;gBACtD,OAAO,CAAC,OAAO,CAAC,SAAS,GAAG,CAAC,CAAC;YAChC,CAAC;QACH,CAAC;QAED,IAAI,IAAI,CAAC,MAAM,CAAC,aAAa,EAAE,CAAC;YAC9B,KAAK,MAAM,OAAO,IAAI,uBAAY,EAAE,CAAC;gBACnC,OAAO,CAAC,OAAO,CAAC,SAAS,GAAG,CAAC,CAAC;gBAC9B,IAAI,OAAO,CAAC,OAAO,CAAC,IAAI,CAAC,KAAK,CAAC,EAAE,CAAC;oBAChC,OAAO,CAAC,OAAO,CAAC,SAAS,GAAG,CAAC,CAAC;oBAC9B,MAAM,KAAK,GAAG,OAAO,CAAC,OAAO,CAAC,IAAI,CAAC,KAAK,CAAC,CAAC;oBAC1C,QAAQ,CAAC,IAAI,CAAC,OAAO,CAAC,IAAI,CAAC,CAAC;oBAC5B,UAAU,CAAC,IAAI,CAAC;wBACd,IAAI;wBACJ,MAAM,EAAE,IAAI,CAAC,MAAM,CAAC,wBAAwB;wBAC5C,aAAa,EAAE,OAAO,CAAC,IAAI;wBAC3B,cAAc,EAAE,KAAK,CAAC,CAAC,CAAC,UAAU,CAAC,SAAS,CAAC,KAAK,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,SAAS;qBACnE,CAAC,CAAC;gBACL,CAAC;gBACD,OAAO,CAAC,OAAO,CAAC,SAAS,GAAG,CAAC,CAAC;YAChC,CAAC;QACH,CAAC;QAED,OAAO,EAAE,QAAQ,EAAE,UAAU,EAAE,CAAC;IAClC,CAAC;IAED,sEAAsE;IACtE,MAAM,CAAC,SAAS,CAAC,KAAa;QAC5B,IAAI,KAAK,CAAC,MAAM,IAAI,CAAC;YAAE,OAAO,KAAK,CAAC,KAAK,CAAC,CAAC,EAAE,CAAC,CAAC,GAAG,KAAK,GAAG,KAAK,CAAC,KAAK,CAAC,CAAC,CAAC,CAAC,CAAC;QAC1E,OAAO,KAAK,CAAC,KAAK,CAAC,CAAC,EAAE,CAAC,CAAC,GAAG,KAAK,GAAG,KAAK,CAAC,KAAK,CAAC,CAAC,CAAC,CAAC,CAAC;IACrD,CAAC;IAED;;;;OAIG;IACK,UAAU,CAChB,GAAY,EACZ,IAAY,EACZ,QAAkB,EAClB,UAA0B,EAC1B,KAAa;QAEb,4DAA4D;QAC5D,IAAI,KAAK,GAAG,cAAc,EAAE,CAAC;YAC3B,OAAO;QACT,CAAC;QAED,IAAI,GAAG,KAAK,IAAI,IAAI,GAAG,KAAK,SAAS,EAAE,CAAC;YACtC,OAAO;QACT,CAAC;QAED,IAAI,OAAO,GAAG,KAAK,QAAQ,EAAE,CAAC;YAC5B,MAAM,MAAM,GAAG,IAAI,CAAC,UAAU,CAAC,GAAG,EAAE,IAAI,CAAC,CAAC;YAC1C,QAAQ,CAAC,IAAI,CAAC,GAAG,MAAM,CAAC,QAAQ,CAAC,CAAC;YAClC,UAAU,CAAC,IAAI,CAAC,GAAG,MAAM,CAAC,UAAU,CAAC,CAAC;YACtC,OAAO;QACT,CAAC;QAED,IAAI,KAAK,CAAC,OAAO,CAAC,GAAG,CAAC,EAAE,CAAC;YACvB,KAAK,IAAI,KAAK,GAAG,CAAC,EAAE,KAAK,GAAG,GAAG,CAAC,MAAM,EAAE,KAAK,EAAE,EAAE,CAAC;gBAChD,MAAM,SAAS,GAAG,IAAI,CAAC,CAAC,CAAC,GAAG,IAAI,IAAI,KAAK,GAAG,CAAC,CAAC,CAAC,IAAI,KAAK,GAAG,CAAC;gBAC5D,IAAI,CAAC,UAAU,CAAC,GAAG,CAAC,KAAK,CAAC,EAAE,SAAS,EAAE,QAAQ,EAAE,UAAU,EAAE,KAAK,GAAG,CAAC,CAAC,CAAC;YAC1E,CAAC;YACD,OAAO;QACT,CAAC;QAED,IAAI,OAAO,GAAG,KAAK,QAAQ,EAAE,CAAC;YAC5B,KAAK,MAAM,CAAC,GAAG,EAAE,KAAK,CAAC,IAAI,MAAM,CAAC,OAAO,CAAC,GAA8B,CAAC,EAAE,CAAC;gBAC1E,MAAM,SAAS,GAAG,IAAI,CAAC,CAAC,CAAC,GAAG,IAAI,IAAI,GAAG,EAAE,CAAC,CAAC,CAAC,GAAG,CAAC;gBAChD,IAAI,CAAC,UAAU,CAAC,KAAK,EAAE,SAAS,EAAE,QAAQ,EAAE,UAAU,EAAE,KAAK,GAAG,CAAC,CAAC,CAAC;YACrE,CAAC;QACH,CAAC;QACD,iFAAiF;IACnF,CAAC;IAED;;;;;;;OAOG;IACH,MAAM,CAAC,MAAM,CAAC,KAAa,EAAE,MAAuB;QAClD,QAAQ,MAAM,EAAE,CAAC;YACf,KAAK,MAAM;gBACT,IAAI,KAAK,CAAC,MAAM,IAAI,CAAC;oBAAE,OAAO,MAAM,CAAC;gBACrC,OAAO,KAAK,CAAC,KAAK,CAAC,CAAC,EAAE,CAAC,CAAC,GAAG,GAAG,CAAC,MAAM,CAAC,IAAI,CAAC,GAAG,CAAC,KAAK,CAAC,MAAM,GAAG,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC;YACxE,KAAK,MAAM;gBACT,OAAO,OAAO,GAAG,IAAA,mBAAU,EAAC,QAAQ,CAAC,CAAC,MAAM,CAAC,KAAK,CAAC,CAAC,MAAM,CAAC,KAAK,CAAC,CAAC,KAAK,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC;YACjF,KAAK,MAAM;gBACT,OAAO,YAAY,CAAC;YACtB,KAAK,UAAU;gBACb,OAAO,aAAa,GAAG,IAAA,mBAAU,EAAC,QAAQ,CAAC,CAAC,MAAM,CAAC,KAAK,CAAC,CAAC,MAAM,CAAC,KAAK,CAAC,CAAC,KAAK,CAAC,CAAC,EAAE,EAAE,CAAC,GAAG,IAAI,CAAC;YAC9F;gBACE,OAAO,YAAY,CAAC;QACxB,CAAC;IACH,CAAC;IAED;;;;;;;;;;;;OAYG;IACH,eAAe,CAAC,IAAa,EAAE,UAA0B;QACvD,IAAI,UAAU,CAAC,MAAM,KAAK,CAAC,EAAE,CAAC;YAC5B,OAAO,IAAI,CAAC;QACd,CAAC;QAED,iDAAiD;QACjD,MAAM,MAAM,GAAG,UAAU,CAAC,SAAS,CAAC,IAAI,CAAC,CAAC;QAE1C,KAAK,MAAM,SAAS,IAAI,UAAU,EAAE,CAAC;YACnC,IAAI,CAAC,oBAAoB,CAAC,MAAM,EAAE,SAAS,CAAC,IAAI,EAAE,SAAS,CAAC,MAAM,CAAC,CAAC;QACtE,CAAC;QAED,OAAO,MAAM,CAAC;IAChB,CAAC;IAED;;;OAGG;IACK,oBAAoB,CAAC,IAAa,EAAE,IAAY,EAAE,MAAuB;QAC/E,MAAM,QAAQ,GAAG,UAAU,CAAC,SAAS,CAAC,IAAI,CAAC,CAAC;QAC5C,IAAI,QAAQ,CAAC,MAAM,KAAK,CAAC,EAAE,CAAC;YAC1B,OAAO;QACT,CAAC;QAED,mCAAmC;QACnC,IAAI,OAAO,GAAY,IAAI,CAAC;QAC5B,KAAK,IAAI,CAAC,GAAG,CAAC,EAAE,CAAC,GAAG,QAAQ,CAAC,MAAM,GAAG,CAAC,EAAE,CAAC,EAAE,EAAE,CAAC;YAC7C,OAAO,GAAG,UAAU,CAAC,QAAQ,CAAC,OAAO,EAAE,QAAQ,CAAC,CAAC,CAAC,CAAC,CAAC;YACpD,IAAI,OAAO,KAAK,IAAI,IAAI,OAAO,KAAK,SAAS,EAAE,CAAC;gBAC9C,OAAO,CAAC,sBAAsB;YAChC,CAAC;QACH,CAAC;QAED,MAAM,WAAW,GAAG,QAAQ,CAAC,QAAQ,CAAC,MAAM,GAAG,CAAC,CAAC,CAAC;QAClD,MAAM,SAAS,GAAG,UAAU,CAAC,QAAQ,CAAC,OAAO,EAAE,WAAW,CAAC,CAAC;QAE5D,IAAI,OAAO,SAAS,KAAK,QAAQ,EAAE,CAAC;YAClC,OAAO,CAAC,oBAAoB;QAC9B,CAAC;QAED,gEAAgE;QAChE,MAAM,QAAQ,GAAG,IAAI,CAAC,qBAAqB,CAAC,SAAS,EAAE,MAAM,CAAC,CAAC;QAC/D,UAAU,CAAC,QAAQ,CAAC,OAAO,EAAE,WAAW,EAAE,QAAQ,CAAC,CAAC;IACtD,CAAC;IAED;;;;;OAKG;IACK,qBAAqB,CAAC,KAAa,EAAE,MAAuB;QAClE,2DAA2D;QAC3D,MAAM,WAAW,GAAa,EAAE,CAAC;QACjC,IAAI,IAAI,CAAC,MAAM,CAAC,iBAAiB,EAAE,CAAC;YAClC,KAAK,MAAM,CAAC,IAAI,0BAAe,EAAE,CAAC;gBAChC,WAAW,CAAC,IAAI,CAAC,CAAC,CAAC,OAAO,CAAC,CAAC;YAC9B,CAAC;QACH,CAAC;QACD,IAAI,IAAI,CAAC,MAAM,CAAC,aAAa,EAAE,CAAC;YAC9B,KAAK,MAAM,CAAC,IAAI,uBAAY,EAAE,CAAC;gBAC7B,WAAW,CAAC,IAAI,CAAC,CAAC,CAAC,OAAO,CAAC,CAAC;YAC9B,CAAC;QACH,CAAC;QAOD,MAAM,OAAO,GAAiB,EAAE,CAAC;QAEjC,KAAK,MAAM,KAAK,IAAI,WAAW,EAAE,CAAC;YAChC,KAAK,CAAC,SAAS,GAAG,CAAC,CAAC;YACpB,IAAI,CAAyB,CAAC;YAC9B,OAAO,CAAC,CAAC,GAAG,KAAK,CAAC,IAAI,CAAC,KAAK,CAAC,CAAC,KAAK,IAAI,EAAE,CAAC;gBACxC,OAAO,CAAC,IAAI,CAAC,EAAE,KAAK,EAAE,CAAC,CAAC,KAAK,EAAE,GAAG,EAAE,CAAC,CAAC,KAAK,GAAG,CAAC,CAAC,CAAC,CAAC,CAAC,MAAM,EAAE,IAAI,EAAE,CAAC,CAAC,CAAC,CAAC,EAAE,CAAC,CAAC;gBACzE,gDAAgD;gBAChD,IAAI,CAAC,CAAC,CAAC,CAAC,CAAC,MAAM,KAAK,CAAC,EAAE,CAAC;oBACtB,KAAK,CAAC,SAAS,EAAE,CAAC;gBACpB,CAAC;YACH,CAAC;YACD,KAAK,CAAC,SAAS,GAAG,CAAC,CAAC;QACtB,CAAC;QAED,IAAI,OAAO,CAAC,MAAM,KAAK,CAAC,EAAE,CAAC;YACzB,oEAAoE;YACpE,OAAO,UAAU,CAAC,MAAM,CAAC,KAAK,EAAE,MAAM,CAAC,CAAC;QAC1C,CAAC;QAED,oFAAoF;QACpF,OAAO,CAAC,IAAI,CAAC,CAAC,CAAC,EAAE,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC,KAAK,GAAG,CAAC,CAAC,KAAK,IAAI,CAAC,CAAC,GAAG,GAAG,CAAC,CAAC,GAAG,CAAC,CAAC;QAE3D,yDAAyD;QACzD,IAAI,MAAM,GAAG,EAAE,CAAC;QAChB,IAAI,MAAM,GAAG,CAAC,CAAC;QAEf,KAAK,MAAM,KAAK,IAAI,OAAO,EAAE,CAAC;YAC5B,0DAA0D;YAC1D,IAAI,KAAK,CAAC,KAAK,GAAG,MAAM,EAAE,CAAC;gBACzB,SAAS;YACX,CAAC;YACD,kDAAkD;YAClD,MAAM,IAAI,KAAK,CAAC,KAAK,CAAC,MAAM,EAAE,KAAK,CAAC,KAAK,CAAC,CAAC;YAC3C,4BAA4B;YAC5B,MAAM,IAAI,UAAU,CAAC,MAAM,CAAC,KAAK,CAAC,IAAI,EAAE,MAAM,CAAC,CAAC;YAChD,MAAM,GAAG,KAAK,CAAC,GAAG,CAAC;QACrB,CAAC;QAED,iDAAiD;QACjD,MAAM,IAAI,KAAK,CAAC,KAAK,CAAC,MAAM,CAAC,CAAC;QAE9B,OAAO,MAAM,CAAC;IAChB,CAAC;IAED;;;;;;;OAOG;IACH,MAAM,CAAC,SAAS,CAAC,IAAY;QAC3B,IAAI,CAAC,IAAI;YAAE,OAAO,EAAE,CAAC;QAErB,MAAM,QAAQ,GAAa,EAAE,CAAC;QAC9B,IAAI,OAAO,GAAG,EAAE,CAAC;QAEjB,KAAK,IAAI,CAAC,GAAG,CAAC,EAAE,CAAC,GAAG,IAAI,CAAC,MAAM,EAAE,CAAC,EAAE,EAAE,CAAC;YACrC,MAAM,IAAI,GAAG,IAAI,CAAC,CAAC,CAAC,CAAC;YACrB,IAAI,IAAI,KAAK,GAAG,EAAE,CAAC;gBACjB,IAAI,OAAO,EAAE,CAAC;oBACZ,QAAQ,CAAC,IAAI,CAAC,OAAO,CAAC,CAAC;oBACvB,OAAO,GAAG,EAAE,CAAC;gBACf,CAAC;YACH,CAAC;iBAAM,IAAI,IAAI,KAAK,GAAG,EAAE,CAAC;gBACxB,IAAI,OAAO,EAAE,CAAC;oBACZ,QAAQ,CAAC,IAAI,CAAC,OAAO,CAAC,CAAC;oBACvB,OAAO,GAAG,EAAE,CAAC;gBACf,CAAC;gBACD,6BAA6B;gBAC7B,MAAM,YAAY,GAAG,IAAI,CAAC,OAAO,CAAC,GAAG,EAAE,CAAC,GAAG,CAAC,CAAC,CAAC;gBAC9C,IAAI,YAAY,KAAK,CAAC,CAAC,EAAE,CAAC;oBACxB,mDAAmD;oBACnD,OAAO,GAAG,IAAI,CAAC,KAAK,CAAC,CAAC,CAAC,CAAC;oBACxB,MAAM;gBACR,CAAC;gBACD,QAAQ,CAAC,IAAI,CAAC,IAAI,CAAC,KAAK,CAAC,CAAC,GAAG,CAAC,EAAE,YAAY,CAAC,CAAC,CAAC;gBAC/C,CAAC,GAAG,YAAY,CAAC,CAAC,mCAAmC;YACvD,CAAC;iBAAM,CAAC;gBACN,OAAO,IAAI,IAAI,CAAC;YAClB,CAAC;QACH,CAAC;QAED,IAAI,OAAO,EAAE,CAAC;YACZ,QAAQ,CAAC,IAAI,CAAC,OAAO,CAAC,CAAC;QACzB,CAAC;QAED,OAAO,QAAQ,CAAC;IAClB,CAAC;IAED;;OAEG;IACK,MAAM,CAAC,QAAQ,CAAC,GAAY,EAAE,OAAe;QACnD,IAAI,GAAG,KAAK,IAAI,IAAI,GAAG,KAAK,SAAS;YAAE,OAAO,SAAS,CAAC;QAExD,IAAI,KAAK,CAAC,OAAO,CAAC,GAAG,CAAC,EAAE,CAAC;YACvB,MAAM,KAAK,GAAG,QAAQ,CAAC,OAAO,EAAE,EAAE,CAAC,CAAC;YACpC,IAAI,MAAM,CAAC,KAAK,CAAC,KAAK,CAAC;gBAAE,OAAO,SAAS,CAAC;YAC1C,OAAO,GAAG,CAAC,KAAK,CAAC,CAAC;QACpB,CAAC;QAED,IAAI,OAAO,GAAG,KAAK,QAAQ,EAAE,CAAC;YAC5B,OAAQ,GAA+B,CAAC,OAAO,CAAC,CAAC;QACnD,CAAC;QAED,OAAO,SAAS,CAAC;IACnB,CAAC;IAED;;OAEG;IACK,MAAM,CAAC,QAAQ,CAAC,GAAY,EAAE,OAAe,EAAE,KAAc;QACnE,IAAI,GAAG,KAAK,IAAI,IAAI,GAAG,KAAK,SAAS;YAAE,OAAO;QAE9C,IAAI,KAAK,CAAC,OAAO,CAAC,GAAG,CAAC,EAAE,CAAC;YACvB,MAAM,KAAK,GAAG,QAAQ,CAAC,OAAO,EAAE,EAAE,CAAC,CAAC;YACpC,IAAI,CAAC,MAAM,CAAC,KAAK,CAAC,KAAK,CAAC,IAAI,KAAK,IAAI,CAAC,IAAI,KAAK,GAAG,GAAG,CAAC,MAAM,EAAE,CAAC;gBAC7D,GAAG,CAAC,KAAK,CAAC,GAAG,KAAK,CAAC;YACrB,CAAC;YACD,OAAO;QACT,CAAC;QAED,IAAI,OAAO,GAAG,KAAK,QAAQ,EAAE,CAAC;YAC3B,GAA+B,CAAC,OAAO,CAAC,GAAG,KAAK,CAAC;QACpD,CAAC;IACH,CAAC;IAED;;;;;OAKG;IACH,MAAM,CAAC,SAAS,CAAI,IAAO;QACzB,IAAI,IAAI,KAAK,IAAI,IAAI,IAAI,KAAK,SAAS;YAAE,OAAO,IAAI,CAAC;QACrD,IAAI,OAAO,IAAI,KAAK,QAAQ;YAAE,OAAO,IAAI,CAAC;QAE1C,gDAAgD;QAChD,IAAI,OAAO,eAAe,KAAK,UAAU,EAAE,CAAC;YAC1C,IAAI,CAAC;gBACH,OAAO,eAAe,CAAC,IAAI,CAAC,CAAC;YAC/B,CAAC;YAAC,MAAM,CAAC;gBACP,4DAA4D;gBAC5D,6CAA6C;YAC/C,CAAC;QACH,CAAC;QAED,IAAI,CAAC;YACH,OAAO,IAAI,CAAC,KAAK,CAAC,IAAI,CAAC,SAAS,CAAC,IAAI,CAAC,CAAC,CAAC;QAC1C,CAAC;QAAC,MAAM,CAAC;YACP,6EAA6E;YAC7E,OAAO,IAAI,CAAC;QACd,CAAC;IACH,CAAC;IAED;;;;;;OAMG;IACK,iBAAiB,CAAC,QAAkB;QAC1C,IAAI,QAAQ,CAAC,MAAM,KAAK,CAAC;YAAE,OAAO,KAAK,CAAC;QAExC,MAAM,WAAW,GAAG,CAAC,GAAG,0BAAe,EAAE,GAAG,uBAAY,CAAC,CAAC;QAE1D,+CAA+C;QAC/C,MAAM,WAAW,GAAG,IAAI,GAAG,EAAuB,CAAC;QACnD,KAAK,MAAM,CAAC,IAAI,WAAW,EAAE,CAAC;YAC5B,WAAW,CAAC,GAAG,CAAC,CAAC,CAAC,IAAI,EAAE,CAAC,CAAC,QAAQ,CAAC,CAAC;QACtC,CAAC;QAED,MAAM,YAAY,GAAgC;YAChD,GAAG,EAAE,CAAC;YACN,MAAM,EAAE,CAAC;YACT,IAAI,EAAE,CAAC;SACR,CAAC;QAEF,IAAI,OAAO,GAAgB,KAAK,CAAC;QAEjC,KAAK,MAAM,IAAI,IAAI,QAAQ,EAAE,CAAC;YAC5B,MAAM,GAAG,GAAG,WAAW,CAAC,GAAG,CAAC,IAAI,CAAC,CAAC;YAClC,IAAI,GAAG,IAAI,YAAY,CAAC,GAAG,CAAC,GAAG,YAAY,CAAC,OAAO,CAAC,EAAE,CAAC;gBACrD,OAAO,GAAG,GAAG,CAAC;YAChB,CAAC;YACD,8CAA8C;YAC9C,IAAI,OAAO,KAAK,MAAM;gBAAE,MAAM;QAChC,CAAC;QAED,OAAO,OAAO,CAAC;IACjB,CAAC;IAED;;;;;;;OAOG;IACH,MAAM,CAAC,gBAAgB,CAAC,GAAW;QACjC,IAAI,GAAG,CAAC,MAAM,KAAK,CAAC;YAAE,OAAO,CAAC,CAAC;QAE/B,MAAM,IAAI,GAA2B,EAAE,CAAC;QACxC,KAAK,MAAM,CAAC,IAAI,GAAG,EAAE,CAAC;YACpB,IAAI,CAAC,CAAC,CAAC,GAAG,CAAC,IAAI,CAAC,CAAC,CAAC,IAAI,CAAC,CAAC,GAAG,CAAC,CAAC;QAC/B,CAAC;QAED,IAAI,OAAO,GAAG,CAAC,CAAC;QAChB,MAAM,GAAG,GAAG,GAAG,CAAC,MAAM,CAAC;QACvB,KAAK,MAAM,KAAK,IAAI,MAAM,CAAC,MAAM,CAAC,IAAI,CAAC,EAAE,CAAC;YACxC,MAAM,CAAC,GAAG,KAAK,GAAG,GAAG,CAAC;YACtB,OAAO,IAAI,CAAC,GAAG,IAAI,CAAC,IAAI,CAAC,CAAC,CAAC,CAAC;QAC9B,CAAC;QAED,OAAO,OAAO,CAAC;IACjB,CAAC;IAED;;;;;;;;;;OAUG;IACH,MAAM,CAAC,aAAa,CAAC,GAAW,EAAE,YAAoB,GAAG;QACvD,IAAI,GAAG,CAAC,MAAM,GAAG,EAAE;YAAE,OAAO,KAAK,CAAC;QAClC,OAAO,UAAU,CAAC,gBAAgB,CAAC,GAAG,CAAC,IAAI,SAAS,CAAC;IACvD,CAAC;CACF;AAzeD,gCAyeC"}

package/dist/src/dlp/text-normalizer.d.ts

@@ -0,0 +1,41 @@
+/**
+ * Text normalizer for bypass-resistant prompt injection detection.
+ *
+ * Applies a series of transformations to collapse evasion techniques
+ * (zero-width chars, HTML entities, homoglyphs, leetspeak, etc.)
+ * into canonical ASCII text before pattern matching.
+ */
+/** Regex matching zero-width and invisible Unicode characters. */
+export declare const ZERO_WIDTH_REGEX: RegExp;
+/** Map of Unicode homoglyphs to their ASCII equivalents. */
+export declare const HOMOGLYPH_MAP: Record<string, string>;
+/** Map of common leetspeak substitutions to their letter equivalents. */
+export declare const LEETSPEAK_MAP: Record<string, string>;
+/**
+ * Normalize text for bypass-resistant pattern matching.
+ *
+ * Applies transformations in order:
+ * 1. Strip zero-width / invisible Unicode characters
+ * 2. Unicode NFKC normalization (collapses fullwidth, ligatures, etc.)
+ * 3. Decode HTML entities (named + numeric)
+ * 4. Decode URL percent-encoding
+ * 5. Collapse homoglyphs (Cyrillic/Greek lookalikes -> ASCII)
+ * 6. Collapse repeated whitespace to single space
+ *
+ * @param input - The raw text to normalize.
+ * @returns The normalized text suitable for pattern matching.
+ */
+export declare function normalizeText(input: string): string;
+/**
+ * Apply leetspeak normalization on top of standard normalization.
+ *
+ * Returns the leet-decoded version of the text. Callers should match
+ * patterns against BOTH the standard-normalized and leet-normalized text
+ * to catch leet evasions without causing false positives on normal text
+ * containing digits.
+ *
+ * @param normalizedInput - Text already passed through normalizeText().
+ * @returns The leet-decoded text.
+ */
+export declare function normalizeLeetspeak(normalizedInput: string): string;
+//# sourceMappingURL=text-normalizer.d.ts.map

package/dist/src/dlp/text-normalizer.d.ts.map

@@ -0,0 +1 @@
+{"version":3,"file":"text-normalizer.d.ts","sourceRoot":"","sources":["../../../src/dlp/text-normalizer.ts"],"names":[],"mappings":"AAAA;;;;;;GAMG;AAMH,kEAAkE;AAClE,eAAO,MAAM,gBAAgB,QAAsF,CAAC;AAMpH,4DAA4D;AAC5D,eAAO,MAAM,aAAa,EAAE,MAAM,CAAC,MAAM,EAAE,MAAM,CAyDhD,CAAC;AAYF,yEAAyE;AACzE,eAAO,MAAM,aAAa,EAAE,MAAM,CAAC,MAAM,EAAE,MAAM,CAUhD,CAAC;AAsEF;;;;;;;;;;;;;GAaG;AACH,wBAAgB,aAAa,CAAC,KAAK,EAAE,MAAM,GAAG,MAAM,CAyBnD;AAED;;;;;;;;;;GAUG;AACH,wBAAgB,kBAAkB,CAAC,eAAe,EAAE,MAAM,GAAG,MAAM,CAElE"}