anantham-v2 2.0.0-alpha.1

package/dist/content/content-ingestion-engine.js

@@ -0,0 +1,164 @@
+import { createHash } from "node:crypto";
+import { ContentObjectSchema, } from "../domain/content.js";
+import { ContentGuards } from "./content-guards.js";
+import { TextParser } from "./parsers/text-parser.js";
+import { StructuredDataParser } from "./parsers/structured-parser.js";
+import { PdfParser } from "./parsers/pdf-parser.js";
+import { ImageParser } from "./parsers/image-parser.js";
+import { MediaParser } from "./parsers/media-parser.js";
+import { ArchiveParser } from "./parsers/archive-parser.js";
+import { BinaryParser } from "./parsers/binary-parser.js";
+export class ContentIngestionEngine {
+    /**
+     * Ingests multimodal content, performs security validation, extracts structured representations,
+     * binds provenance, and produces a validated, immutable ContentObject.
+     * PRD Part 1 Section 10-13 & PRD Part 3 Section 135-140.
+     */
+    static async ingest(request) {
+        const buffer = Buffer.isBuffer(request.data)
+            ? request.data
+            : Buffer.from(request.data, "utf8");
+        // 1. Enforce Size Guard
+        ContentGuards.validateSize(buffer.length, request.maxSizeBytes);
+        // 2. Compute Deterministic SHA-256 Digest
+        const sha256 = createHash("sha256").update(buffer).digest("hex");
+        // 3. Sniff MIME Type and Classify Kind
+        const sniffResult = ContentGuards.sniffMimeType(buffer, request.name);
+        const mimeType = sniffResult.mimeType;
+        let kind = "binary";
+        if (mimeType === "application/json" ||
+            mimeType === "text/csv" ||
+            mimeType === "text/tab-separated-values") {
+            kind = "table";
+        }
+        else if (mimeType.startsWith("text/code") ||
+            mimeType.includes("javascript") ||
+            mimeType.includes("typescript") ||
+            mimeType.includes("python") ||
+            mimeType.includes("rust") ||
+            mimeType.includes("golang") ||
+            mimeType.includes("sql")) {
+            kind = "code";
+        }
+        else if (mimeType.startsWith("text/")) {
+            kind = "text";
+        }
+        else if (mimeType === "application/pdf" ||
+            mimeType.includes("wordprocessingml") ||
+            mimeType.includes("spreadsheetml")) {
+            kind = "document";
+        }
+        else if (mimeType.startsWith("image/")) {
+            kind = "image";
+        }
+        else if (mimeType.startsWith("audio/")) {
+            kind = "audio";
+        }
+        else if (mimeType.startsWith("video/")) {
+            kind = "video";
+        }
+        else if (mimeType === "application/zip" ||
+            mimeType === "application/gzip" ||
+            mimeType.includes("tar")) {
+            kind = "archive";
+        }
+        // 4. Dispatch to Specialized Parser
+        let representations = [];
+        switch (kind) {
+            case "text":
+            case "code": {
+                const res = TextParser.parse(buffer, request.name, mimeType);
+                representations = res.representations;
+                break;
+            }
+            case "table": {
+                const format = mimeType === "text/csv"
+                    ? "csv"
+                    : mimeType === "text/tab-separated-values"
+                        ? "tsv"
+                        : "json";
+                const res = StructuredDataParser.parse(buffer, format);
+                representations = res.representations;
+                break;
+            }
+            case "document": {
+                if (mimeType === "application/pdf") {
+                    const res = PdfParser.parse(buffer);
+                    representations = res.representations;
+                }
+                else {
+                    const res = BinaryParser.parse(buffer, mimeType);
+                    representations = res.representations;
+                }
+                break;
+            }
+            case "image": {
+                const res = ImageParser.parse(buffer, mimeType);
+                representations = res.representations;
+                break;
+            }
+            case "audio":
+            case "video": {
+                const res = MediaParser.parse(buffer, mimeType);
+                representations = res.representations;
+                break;
+            }
+            case "archive": {
+                const res = ArchiveParser.parse(buffer, mimeType);
+                representations = res.representations;
+                break;
+            }
+            default: {
+                const res = BinaryParser.parse(buffer, mimeType);
+                representations = res.representations;
+                break;
+            }
+        }
+        // 5. Construct ContentObject with Provenance and Security
+        const id = `cnt_${sha256.slice(0, 16)}`;
+        const now = new Date().toISOString();
+        const trust = request.trust ||
+            (request.source.type === "upload" || request.source.type === "clipboard"
+                ? "user-content"
+                : request.source.type === "filesystem"
+                    ? "repository-content"
+                    : request.source.type === "browser"
+                        ? "web-content"
+                        : request.source.type === "mcp"
+                            ? "mcp-content"
+                            : "trusted");
+        const contentObject = {
+            id,
+            kind,
+            mimeType,
+            name: request.name,
+            sizeBytes: buffer.length,
+            sha256,
+            source: request.source,
+            representations,
+            provenance: {
+                sourceType: request.source.type,
+                sourceId: request.actor,
+                sourceUri: request.source.uri,
+                parentIds: [],
+                capturedAt: now,
+                extractor: {
+                    name: "content-ingestion-engine",
+                    version: "2.0.0",
+                },
+                transformations: [`parsed:${kind}`],
+            },
+            security: {
+                trust,
+                sensitivity: request.sensitivity || "normal",
+                scanned: true,
+                scanVersion: "1.0.0",
+                authority: request.source.type === "upload" ? "attachment" : "tool-output",
+            },
+            createdAt: now,
+            updatedAt: now,
+        };
+        return Object.freeze(ContentObjectSchema.parse(contentObject));
+    }
+}
+//# sourceMappingURL=content-ingestion-engine.js.map

package/dist/content/content-ingestion-engine.js.map

@@ -0,0 +1 @@
+{"version":3,"file":"content-ingestion-engine.js","sourceRoot":"","sources":["../../src/content/content-ingestion-engine.ts"],"names":[],"mappings":"AAAA,OAAO,EAAE,UAAU,EAAE,MAAM,aAAa,CAAC;AACzC,OAAO,EAIL,mBAAmB,GACpB,MAAM,sBAAsB,CAAC;AAE9B,OAAO,EAAE,aAAa,EAAE,MAAM,qBAAqB,CAAC;AACpD,OAAO,EAAE,UAAU,EAAE,MAAM,0BAA0B,CAAC;AACtD,OAAO,EAAE,oBAAoB,EAAE,MAAM,gCAAgC,CAAC;AACtE,OAAO,EAAE,SAAS,EAAE,MAAM,yBAAyB,CAAC;AACpD,OAAO,EAAE,WAAW,EAAE,MAAM,2BAA2B,CAAC;AACxD,OAAO,EAAE,WAAW,EAAE,MAAM,2BAA2B,CAAC;AACxD,OAAO,EAAE,aAAa,EAAE,MAAM,6BAA6B,CAAC;AAC5D,OAAO,EAAE,YAAY,EAAE,MAAM,4BAA4B,CAAC;AAe1D,MAAM,OAAO,sBAAsB;IACjC;;;;OAIG;IACI,MAAM,CAAC,KAAK,CAAC,MAAM,CAAC,OAA6B;QACtD,MAAM,MAAM,GAAG,MAAM,CAAC,QAAQ,CAAC,OAAO,CAAC,IAAI,CAAC;YAC1C,CAAC,CAAC,OAAO,CAAC,IAAI;YACd,CAAC,CAAC,MAAM,CAAC,IAAI,CAAC,OAAO,CAAC,IAAI,EAAE,MAAM,CAAC,CAAC;QAEtC,wBAAwB;QACxB,aAAa,CAAC,YAAY,CAAC,MAAM,CAAC,MAAM,EAAE,OAAO,CAAC,YAAY,CAAC,CAAC;QAEhE,0CAA0C;QAC1C,MAAM,MAAM,GAAG,UAAU,CAAC,QAAQ,CAAC,CAAC,MAAM,CAAC,MAAM,CAAC,CAAC,MAAM,CAAC,KAAK,CAAC,CAAC;QAEjE,uCAAuC;QACvC,MAAM,WAAW,GAAG,aAAa,CAAC,aAAa,CAAC,MAAM,EAAE,OAAO,CAAC,IAAI,CAAC,CAAC;QACtE,MAAM,QAAQ,GAAG,WAAW,CAAC,QAAQ,CAAC;QAEtC,IAAI,IAAI,GAAgB,QAAQ,CAAC;QACjC,IACE,QAAQ,KAAK,kBAAkB;YAC/B,QAAQ,KAAK,UAAU;YACvB,QAAQ,KAAK,2BAA2B,EACxC,CAAC;YACD,IAAI,GAAG,OAAO,CAAC;QACjB,CAAC;aAAM,IACL,QAAQ,CAAC,UAAU,CAAC,WAAW,CAAC;YAChC,QAAQ,CAAC,QAAQ,CAAC,YAAY,CAAC;YAC/B,QAAQ,CAAC,QAAQ,CAAC,YAAY,CAAC;YAC/B,QAAQ,CAAC,QAAQ,CAAC,QAAQ,CAAC;YAC3B,QAAQ,CAAC,QAAQ,CAAC,MAAM,CAAC;YACzB,QAAQ,CAAC,QAAQ,CAAC,QAAQ,CAAC;YAC3B,QAAQ,CAAC,QAAQ,CAAC,KAAK,CAAC,EACxB,CAAC;YACD,IAAI,GAAG,MAAM,CAAC;QAChB,CAAC;aAAM,IAAI,QAAQ,CAAC,UAAU,CAAC,OAAO,CAAC,EAAE,CAAC;YACxC,IAAI,GAAG,MAAM,CAAC;QAChB,CAAC;aAAM,IACL,QAAQ,KAAK,iBAAiB;YAC9B,QAAQ,CAAC,QAAQ,CAAC,kBAAkB,CAAC;YACrC,QAAQ,CAAC,QAAQ,CAAC,eAAe,CAAC,EAClC,CAAC;YACD,IAAI,GAAG,UAAU,CAAC;QACpB,CAAC;aAAM,IAAI,QAAQ,CAAC,UAAU,CAAC,QAAQ,CAAC,EAAE,CAAC;YACzC,IAAI,GAAG,OAAO,CAAC;QACjB,CAAC;aAAM,IAAI,QAAQ,CAAC,UAAU,CAAC,QAAQ,CAAC,EAAE,CAAC;YACzC,IAAI,GAAG,OAAO,CAAC;QACjB,CAAC;aAAM,IAAI,QAAQ,CAAC,UAAU,CAAC,QAAQ,CAAC,EAAE,CAAC;YACzC,IAAI,GAAG,OAAO,CAAC;QACjB,CAAC;aAAM,IACL,QAAQ,KAAK,iBAAiB;YAC9B,QAAQ,KAAK,kBAAkB;YAC/B,QAAQ,CAAC,QAAQ,CAAC,KAAK,CAAC,EACxB,CAAC;YACD,IAAI,GAAG,SAAS,CAAC;QACnB,CAAC;QAED,oCAAoC;QACpC,IAAI,eAAe,GAAG,EAAE,CAAC;QAEzB,QAAQ,IAAI,EAAE,CAAC;YACb,KAAK,MAAM,CAAC;YACZ,KAAK,MAAM,CAAC,CAAC,CAAC;gBACZ,MAAM,GAAG,GAAG,UAAU,CAAC,KAAK,CAAC,MAAM,EAAE,OAAO,CAAC,IAAI,EAAE,QAAQ,CAAC,CAAC;gBAC7D,eAAe,GAAG,GAAG,CAAC,eAAe,CAAC;gBACtC,MAAM;YACR,CAAC;YAED,KAAK,OAAO,CAAC,CAAC,CAAC;gBACb,MAAM,MAAM,GACV,QAAQ,KAAK,UAAU;oBACrB,CAAC,CAAC,KAAK;oBACP,CAAC,CAAC,QAAQ,KAAK,2BAA2B;wBAC1C,CAAC,CAAC,KAAK;wBACP,CAAC,CAAC,MAAM,CAAC;gBACb,MAAM,GAAG,GAAG,oBAAoB,CAAC,KAAK,CAAC,MAAM,EAAE,MAAM,CAAC,CAAC;gBACvD,eAAe,GAAG,GAAG,CAAC,eAAe,CAAC;gBACtC,MAAM;YACR,CAAC;YAED,KAAK,UAAU,CAAC,CAAC,CAAC;gBAChB,IAAI,QAAQ,KAAK,iBAAiB,EAAE,CAAC;oBACnC,MAAM,GAAG,GAAG,SAAS,CAAC,KAAK,CAAC,MAAM,CAAC,CAAC;oBACpC,eAAe,GAAG,GAAG,CAAC,eAAe,CAAC;gBACxC,CAAC;qBAAM,CAAC;oBACN,MAAM,GAAG,GAAG,YAAY,CAAC,KAAK,CAAC,MAAM,EAAE,QAAQ,CAAC,CAAC;oBACjD,eAAe,GAAG,GAAG,CAAC,eAAe,CAAC;gBACxC,CAAC;gBACD,MAAM;YACR,CAAC;YAED,KAAK,OAAO,CAAC,CAAC,CAAC;gBACb,MAAM,GAAG,GAAG,WAAW,CAAC,KAAK,CAAC,MAAM,EAAE,QAAQ,CAAC,CAAC;gBAChD,eAAe,GAAG,GAAG,CAAC,eAAe,CAAC;gBACtC,MAAM;YACR,CAAC;YAED,KAAK,OAAO,CAAC;YACb,KAAK,OAAO,CAAC,CAAC,CAAC;gBACb,MAAM,GAAG,GAAG,WAAW,CAAC,KAAK,CAAC,MAAM,EAAE,QAAQ,CAAC,CAAC;gBAChD,eAAe,GAAG,GAAG,CAAC,eAAe,CAAC;gBACtC,MAAM;YACR,CAAC;YAED,KAAK,SAAS,CAAC,CAAC,CAAC;gBACf,MAAM,GAAG,GAAG,aAAa,CAAC,KAAK,CAAC,MAAM,EAAE,QAAQ,CAAC,CAAC;gBAClD,eAAe,GAAG,GAAG,CAAC,eAAe,CAAC;gBACtC,MAAM;YACR,CAAC;YAED,OAAO,CAAC,CAAC,CAAC;gBACR,MAAM,GAAG,GAAG,YAAY,CAAC,KAAK,CAAC,MAAM,EAAE,QAAQ,CAAC,CAAC;gBACjD,eAAe,GAAG,GAAG,CAAC,eAAe,CAAC;gBACtC,MAAM;YACR,CAAC;QACH,CAAC;QAED,0DAA0D;QAC1D,MAAM,EAAE,GAAG,OAAO,MAAM,CAAC,KAAK,CAAC,CAAC,EAAE,EAAE,CAAC,EAAE,CAAC;QACxC,MAAM,GAAG,GAAG,IAAI,IAAI,EAAE,CAAC,WAAW,EAAE,CAAC;QAErC,MAAM,KAAK,GACT,OAAO,CAAC,KAAK;YACb,CAAC,OAAO,CAAC,MAAM,CAAC,IAAI,KAAK,QAAQ,IAAI,OAAO,CAAC,MAAM,CAAC,IAAI,KAAK,WAAW;gBACtE,CAAC,CAAC,cAAc;gBAChB,CAAC,CAAC,OAAO,CAAC,MAAM,CAAC,IAAI,KAAK,YAAY;oBACtC,CAAC,CAAC,oBAAoB;oBACtB,CAAC,CAAC,OAAO,CAAC,MAAM,CAAC,IAAI,KAAK,SAAS;wBACnC,CAAC,CAAC,aAAa;wBACf,CAAC,CAAC,OAAO,CAAC,MAAM,CAAC,IAAI,KAAK,KAAK;4BAC/B,CAAC,CAAC,aAAa;4BACf,CAAC,CAAC,SAAS,CAAC,CAAC;QAEjB,MAAM,aAAa,GAAkB;YACnC,EAAE;YACF,IAAI;YACJ,QAAQ;YACR,IAAI,EAAE,OAAO,CAAC,IAAI;YAClB,SAAS,EAAE,MAAM,CAAC,MAAM;YACxB,MAAM;YACN,MAAM,EAAE,OAAO,CAAC,MAAM;YACtB,eAAe;YACf,UAAU,EAAE;gBACV,UAAU,EAAE,OAAO,CAAC,MAAM,CAAC,IAAI;gBAC/B,QAAQ,EAAE,OAAO,CAAC,KAAK;gBACvB,SAAS,EAAE,OAAO,CAAC,MAAM,CAAC,GAAG;gBAC7B,SAAS,EAAE,EAAE;gBACb,UAAU,EAAE,GAAG;gBACf,SAAS,EAAE;oBACT,IAAI,EAAE,0BAA0B;oBAChC,OAAO,EAAE,OAAO;iBACjB;gBACD,eAAe,EAAE,CAAC,UAAU,IAAI,EAAE,CAAC;aACpC;YACD,QAAQ,EAAE;gBACR,KAAK;gBACL,WAAW,EAAE,OAAO,CAAC,WAAW,IAAI,QAAQ;gBAC5C,OAAO,EAAE,IAAI;gBACb,WAAW,EAAE,OAAO;gBACpB,SAAS,EAAE,OAAO,CAAC,MAAM,CAAC,IAAI,KAAK,QAAQ,CAAC,CAAC,CAAC,YAAY,CAAC,CAAC,CAAC,aAAa;aAC3E;YACD,SAAS,EAAE,GAAG;YACd,SAAS,EAAE,GAAG;SACf,CAAC;QAEF,OAAO,MAAM,CAAC,MAAM,CAAC,mBAAmB,CAAC,KAAK,CAAC,aAAa,CAAC,CAAC,CAAC;IACjE,CAAC;CACF"}

package/dist/content/content-reference-manager.d.ts

@@ -0,0 +1,24 @@
+export interface StoredBlobDescriptor {
+    uri: string;
+    sha256: string;
+    sizeBytes: number;
+}
+export declare class ContentReferenceManager {
+    private static defaultStorageDir;
+    static setStorageDir(dir: string): void;
+    static getStorageDir(): string;
+    /**
+     * Stores a large payload to persistent file storage, returning its SHA-256 digest and URI.
+     * PRD Part 1 Section 10 & PRD Part 3 Section 139.
+     */
+    static storeBlob(data: Buffer, customDir?: string): StoredBlobDescriptor;
+    /**
+     * Reads a stored content blob and strictly verifies its cryptographic SHA-256 checksum.
+     */
+    static readBlob(uri: string, expectedSha256?: string): Buffer;
+    /**
+     * Verifies whether a blob exists and matches the expected SHA-256 hash.
+     */
+    static verifyBlob(uri: string, expectedSha256: string): boolean;
+}
+//# sourceMappingURL=content-reference-manager.d.ts.map

package/dist/content/content-reference-manager.d.ts.map

@@ -0,0 +1 @@
+{"version":3,"file":"content-reference-manager.d.ts","sourceRoot":"","sources":["../../src/content/content-reference-manager.ts"],"names":[],"mappings":"AAIA,MAAM,WAAW,oBAAoB;IACnC,GAAG,EAAE,MAAM,CAAC;IACZ,MAAM,EAAE,MAAM,CAAC;IACf,SAAS,EAAE,MAAM,CAAC;CACnB;AAED,qBAAa,uBAAuB;IAClC,OAAO,CAAC,MAAM,CAAC,iBAAiB,CAAwD;WAE1E,aAAa,CAAC,GAAG,EAAE,MAAM,GAAG,IAAI;WAIhC,aAAa,IAAI,MAAM;IAIrC;;;OAGG;WACW,SAAS,CAAC,IAAI,EAAE,MAAM,EAAE,SAAS,CAAC,EAAE,MAAM,GAAG,oBAAoB;IAkB/E;;OAEG;WACW,QAAQ,CAAC,GAAG,EAAE,MAAM,EAAE,cAAc,CAAC,EAAE,MAAM,GAAG,MAAM;IAsBpE;;OAEG;WACW,UAAU,CAAC,GAAG,EAAE,MAAM,EAAE,cAAc,EAAE,MAAM,GAAG,OAAO;CAQvE"}

package/dist/content/content-reference-manager.js

@@ -0,0 +1,62 @@
+import { createHash } from "node:crypto";
+import { existsSync, mkdirSync, readFileSync, writeFileSync } from "node:fs";
+import { join, resolve } from "node:path";
+export class ContentReferenceManager {
+    static defaultStorageDir = resolve(process.cwd(), ".anantham", "content_blobs");
+    static setStorageDir(dir) {
+        ContentReferenceManager.defaultStorageDir = resolve(dir);
+    }
+    static getStorageDir() {
+        return ContentReferenceManager.defaultStorageDir;
+    }
+    /**
+     * Stores a large payload to persistent file storage, returning its SHA-256 digest and URI.
+     * PRD Part 1 Section 10 & PRD Part 3 Section 139.
+     */
+    static storeBlob(data, customDir) {
+        const sha256 = createHash("sha256").update(data).digest("hex");
+        const targetDir = customDir ? resolve(customDir) : ContentReferenceManager.defaultStorageDir;
+        if (!existsSync(targetDir)) {
+            mkdirSync(targetDir, { recursive: true });
+        }
+        const filePath = join(targetDir, `blob_${sha256}.bin`);
+        writeFileSync(filePath, data);
+        return {
+            uri: `file:///${filePath.replace(/\\/g, "/")}`,
+            sha256,
+            sizeBytes: data.length,
+        };
+    }
+    /**
+     * Reads a stored content blob and strictly verifies its cryptographic SHA-256 checksum.
+     */
+    static readBlob(uri, expectedSha256) {
+        const rawPath = uri.replace(/^file:\/\/\/?/, "");
+        // On Windows handle drive letters properly
+        const filePath = /^[a-zA-Z]:/.test(rawPath) ? rawPath : `/${rawPath}`;
+        if (!existsSync(filePath)) {
+            throw new Error(`Content blob at '${uri}' does not exist on disk.`);
+        }
+        const buffer = readFileSync(filePath);
+        if (expectedSha256) {
+            const actualSha = createHash("sha256").update(buffer).digest("hex");
+            if (actualSha !== expectedSha256) {
+                throw new Error(`Content blob integrity failure: expected SHA-256 '${expectedSha256}', received '${actualSha}'.`);
+            }
+        }
+        return buffer;
+    }
+    /**
+     * Verifies whether a blob exists and matches the expected SHA-256 hash.
+     */
+    static verifyBlob(uri, expectedSha256) {
+        try {
+            ContentReferenceManager.readBlob(uri, expectedSha256);
+            return true;
+        }
+        catch {
+            return false;
+        }
+    }
+}
+//# sourceMappingURL=content-reference-manager.js.map

package/dist/content/content-reference-manager.js.map

@@ -0,0 +1 @@
+{"version":3,"file":"content-reference-manager.js","sourceRoot":"","sources":["../../src/content/content-reference-manager.ts"],"names":[],"mappings":"AAAA,OAAO,EAAE,UAAU,EAAE,MAAM,aAAa,CAAC;AACzC,OAAO,EAAE,UAAU,EAAE,SAAS,EAAE,YAAY,EAAE,aAAa,EAAE,MAAM,SAAS,CAAC;AAC7E,OAAO,EAAE,IAAI,EAAE,OAAO,EAAE,MAAM,WAAW,CAAC;AAQ1C,MAAM,OAAO,uBAAuB;IAC1B,MAAM,CAAC,iBAAiB,GAAG,OAAO,CAAC,OAAO,CAAC,GAAG,EAAE,EAAE,WAAW,EAAE,eAAe,CAAC,CAAC;IAEjF,MAAM,CAAC,aAAa,CAAC,GAAW;QACrC,uBAAuB,CAAC,iBAAiB,GAAG,OAAO,CAAC,GAAG,CAAC,CAAC;IAC3D,CAAC;IAEM,MAAM,CAAC,aAAa;QACzB,OAAO,uBAAuB,CAAC,iBAAiB,CAAC;IACnD,CAAC;IAED;;;OAGG;IACI,MAAM,CAAC,SAAS,CAAC,IAAY,EAAE,SAAkB;QACtD,MAAM,MAAM,GAAG,UAAU,CAAC,QAAQ,CAAC,CAAC,MAAM,CAAC,IAAI,CAAC,CAAC,MAAM,CAAC,KAAK,CAAC,CAAC;QAC/D,MAAM,SAAS,GAAG,SAAS,CAAC,CAAC,CAAC,OAAO,CAAC,SAAS,CAAC,CAAC,CAAC,CAAC,uBAAuB,CAAC,iBAAiB,CAAC;QAE7F,IAAI,CAAC,UAAU,CAAC,SAAS,CAAC,EAAE,CAAC;YAC3B,SAAS,CAAC,SAAS,EAAE,EAAE,SAAS,EAAE,IAAI,EAAE,CAAC,CAAC;QAC5C,CAAC;QAED,MAAM,QAAQ,GAAG,IAAI,CAAC,SAAS,EAAE,QAAQ,MAAM,MAAM,CAAC,CAAC;QACvD,aAAa,CAAC,QAAQ,EAAE,IAAI,CAAC,CAAC;QAE9B,OAAO;YACL,GAAG,EAAE,WAAW,QAAQ,CAAC,OAAO,CAAC,KAAK,EAAE,GAAG,CAAC,EAAE;YAC9C,MAAM;YACN,SAAS,EAAE,IAAI,CAAC,MAAM;SACvB,CAAC;IACJ,CAAC;IAED;;OAEG;IACI,MAAM,CAAC,QAAQ,CAAC,GAAW,EAAE,cAAuB;QACzD,MAAM,OAAO,GAAG,GAAG,CAAC,OAAO,CAAC,eAAe,EAAE,EAAE,CAAC,CAAC;QACjD,2CAA2C;QAC3C,MAAM,QAAQ,GAAG,YAAY,CAAC,IAAI,CAAC,OAAO,CAAC,CAAC,CAAC,CAAC,OAAO,CAAC,CAAC,CAAC,IAAI,OAAO,EAAE,CAAC;QAEtE,IAAI,CAAC,UAAU,CAAC,QAAQ,CAAC,EAAE,CAAC;YAC1B,MAAM,IAAI,KAAK,CAAC,oBAAoB,GAAG,2BAA2B,CAAC,CAAC;QACtE,CAAC;QAED,MAAM,MAAM,GAAG,YAAY,CAAC,QAAQ,CAAC,CAAC;QACtC,IAAI,cAAc,EAAE,CAAC;YACnB,MAAM,SAAS,GAAG,UAAU,CAAC,QAAQ,CAAC,CAAC,MAAM,CAAC,MAAM,CAAC,CAAC,MAAM,CAAC,KAAK,CAAC,CAAC;YACpE,IAAI,SAAS,KAAK,cAAc,EAAE,CAAC;gBACjC,MAAM,IAAI,KAAK,CACb,qDAAqD,cAAc,gBAAgB,SAAS,IAAI,CACjG,CAAC;YACJ,CAAC;QACH,CAAC;QAED,OAAO,MAAM,CAAC;IAChB,CAAC;IAED;;OAEG;IACI,MAAM,CAAC,UAAU,CAAC,GAAW,EAAE,cAAsB;QAC1D,IAAI,CAAC;YACH,uBAAuB,CAAC,QAAQ,CAAC,GAAG,EAAE,cAAc,CAAC,CAAC;YACtD,OAAO,IAAI,CAAC;QACd,CAAC;QAAC,MAAM,CAAC;YACP,OAAO,KAAK,CAAC;QACf,CAAC;IACH,CAAC"}

package/dist/content/content-sanitizer.d.ts

@@ -0,0 +1,39 @@
+import type { ContentObject } from "../domain/content.js";
+export interface SecretFinding {
+    type: "openai-api-key" | "github-pat" | "aws-access-key" | "private-key" | "bearer-token" | "db-credential";
+    matchedPattern: string;
+    index: number;
+}
+export interface SecretDetectionResult {
+    hasSecrets: boolean;
+    findings: SecretFinding[];
+}
+export declare class ContentSanitizer {
+    private static readonly SECRET_PATTERNS;
+    /**
+     * Scans text for exposed credentials and secrets.
+     * PRD Part 3 Section 139.
+     */
+    static scanSecrets(text: string): SecretDetectionResult;
+    /**
+     * Redacts exposed credentials and secrets from text.
+     */
+    static redactSecrets(text: string): {
+        redactedText: string;
+        findings: SecretFinding[];
+    };
+    /**
+     * Sanitizes all textual representations within a ContentObject.
+     * If secrets are found, replaces text with redacted version and escalates sensitivity.
+     */
+    static sanitizeContentObject(content: ContentObject): {
+        sanitized: ContentObject;
+        secretsFound: boolean;
+        findingsCount: number;
+    };
+    /**
+     * Recursively sanitizes any data structure (objects, arrays, strings) for secret-safe logging.
+     */
+    static sanitize(value: unknown): any;
+}
+//# sourceMappingURL=content-sanitizer.d.ts.map

package/dist/content/content-sanitizer.d.ts.map

@@ -0,0 +1 @@
+{"version":3,"file":"content-sanitizer.d.ts","sourceRoot":"","sources":["../../src/content/content-sanitizer.ts"],"names":[],"mappings":"AAAA,OAAO,KAAK,EAAE,aAAa,EAAyB,MAAM,sBAAsB,CAAC;AAGjF,MAAM,WAAW,aAAa;IAC5B,IAAI,EAAE,gBAAgB,GAAG,YAAY,GAAG,gBAAgB,GAAG,aAAa,GAAG,cAAc,GAAG,eAAe,CAAC;IAC5G,cAAc,EAAE,MAAM,CAAC;IACvB,KAAK,EAAE,MAAM,CAAC;CACf;AAED,MAAM,WAAW,qBAAqB;IACpC,UAAU,EAAE,OAAO,CAAC;IACpB,QAAQ,EAAE,aAAa,EAAE,CAAC;CAC3B;AAED,qBAAa,gBAAgB;IAC3B,OAAO,CAAC,MAAM,CAAC,QAAQ,CAAC,eAAe,CAmCrC;IAEF;;;OAGG;WACW,WAAW,CAAC,IAAI,EAAE,MAAM,GAAG,qBAAqB;IAuB9D;;OAEG;WACW,aAAa,CAAC,IAAI,EAAE,MAAM,GAAG;QAAE,YAAY,EAAE,MAAM,CAAC;QAAC,QAAQ,EAAE,aAAa,EAAE,CAAA;KAAE;IAyB9F;;;OAGG;WACW,qBAAqB,CAAC,OAAO,EAAE,aAAa,GAAG;QAC3D,SAAS,EAAE,aAAa,CAAC;QACzB,YAAY,EAAE,OAAO,CAAC;QACtB,aAAa,EAAE,MAAM,CAAC;KACvB;IA2CD;;OAEG;WACW,QAAQ,CAAC,KAAK,EAAE,OAAO,GAAG,GAAG;CA8B5C"}

package/dist/content/content-sanitizer.js

@@ -0,0 +1,155 @@
+import { ContentObjectSchema } from "../domain/content.js";
+export class ContentSanitizer {
+    static SECRET_PATTERNS = [
+        {
+            type: "private-key",
+            regex: /-----BEGIN [A-Z ]*PRIVATE KEY-----[\s\S]*?-----END [A-Z ]*PRIVATE KEY-----/g,
+            redaction: "[REDACTED_PRIVATE_KEY]",
+        },
+        {
+            type: "openai-api-key",
+            regex: /\bsk-[a-zA-Z0-9_\-]{20,}\b/g,
+            redaction: "[REDACTED_OPENAI_API_KEY]",
+        },
+        {
+            type: "github-pat",
+            regex: /\b(?:ghp|gho|ghu|ghs|ghr)_[a-zA-Z0-9]{36}\b|\bgithub_pat_[a-zA-Z0-9_]{40,}\b/g,
+            redaction: "[REDACTED_GITHUB_PAT]",
+        },
+        {
+            type: "aws-access-key",
+            regex: /\bAKIA[0-9A-Z]{16}\b/g,
+            redaction: "[REDACTED_AWS_ACCESS_KEY]",
+        },
+        {
+            type: "bearer-token",
+            regex: /\bBearer\s+[a-zA-Z0-9_\-\.]{25,}\b/gi,
+            redaction: "Bearer [REDACTED_BEARER_TOKEN]",
+        },
+        {
+            type: "db-credential",
+            regex: /\b(?:postgres|postgresql|mongodb|mysql|redis):\/\/[a-zA-Z0-9_\-]+:([^\s@]+)@/g,
+            redaction: "[REDACTED_DB_CREDENTIAL]",
+        },
+    ];
+    /**
+     * Scans text for exposed credentials and secrets.
+     * PRD Part 3 Section 139.
+     */
+    static scanSecrets(text) {
+        if (!text)
+            return { hasSecrets: false, findings: [] };
+        const findings = [];
+        for (const { type, regex } of ContentSanitizer.SECRET_PATTERNS) {
+            regex.lastIndex = 0;
+            let match;
+            while ((match = regex.exec(text)) !== null) {
+                findings.push({
+                    type,
+                    matchedPattern: match[0].slice(0, 8) + "...",
+                    index: match.index,
+                });
+            }
+        }
+        return {
+            hasSecrets: findings.length > 0,
+            findings,
+        };
+    }
+    /**
+     * Redacts exposed credentials and secrets from text.
+     */
+    static redactSecrets(text) {
+        if (!text)
+            return { redactedText: "", findings: [] };
+        let redactedText = text;
+        const findings = [];
+        for (const { type, regex, redaction } of ContentSanitizer.SECRET_PATTERNS) {
+            regex.lastIndex = 0;
+            let match;
+            while ((match = regex.exec(text)) !== null) {
+                findings.push({
+                    type,
+                    matchedPattern: match[0].slice(0, 8) + "...",
+                    index: match.index,
+                });
+            }
+            redactedText = redactedText.replace(regex, redaction);
+        }
+        return {
+            redactedText,
+            findings,
+        };
+    }
+    /**
+     * Sanitizes all textual representations within a ContentObject.
+     * If secrets are found, replaces text with redacted version and escalates sensitivity.
+     */
+    static sanitizeContentObject(content) {
+        let totalFindings = 0;
+        const sanitizedRepresentations = [];
+        for (const rep of content.representations) {
+            if (rep.type === "text" || rep.type === "markdown" || rep.type === "transcript") {
+                if (typeof rep.data === "string") {
+                    const { redactedText, findings } = ContentSanitizer.redactSecrets(rep.data);
+                    totalFindings += findings.length;
+                    sanitizedRepresentations.push({
+                        ...rep,
+                        data: redactedText,
+                    });
+                    continue;
+                }
+            }
+            sanitizedRepresentations.push(rep);
+        }
+        const secretsFound = totalFindings > 0;
+        const sensitivity = secretsFound
+            ? (content.security.sensitivity === "secret" ? "secret" : "sensitive")
+            : content.security.sensitivity;
+        const sanitized = {
+            ...content,
+            representations: sanitizedRepresentations,
+            security: {
+                ...content.security,
+                sensitivity,
+                scanned: true,
+                scanVersion: "2.0.0",
+            },
+            updatedAt: new Date().toISOString(),
+        };
+        return {
+            sanitized: Object.freeze(ContentObjectSchema.parse(sanitized)),
+            secretsFound,
+            findingsCount: totalFindings,
+        };
+    }
+    /**
+     * Recursively sanitizes any data structure (objects, arrays, strings) for secret-safe logging.
+     */
+    static sanitize(value) {
+        if (value === null || value === undefined) {
+            return value;
+        }
+        if (typeof value === "string") {
+            return ContentSanitizer.redactSecrets(value).redactedText;
+        }
+        if (Array.isArray(value)) {
+            return value.map((item) => ContentSanitizer.sanitize(item));
+        }
+        if (typeof value === "object") {
+            const result = {};
+            const SENSITIVE_KEY_REGEX = /(api[_-]?key|password|secret|auth|token|credential|bearer)/i;
+            for (const [k, v] of Object.entries(value)) {
+                if (SENSITIVE_KEY_REGEX.test(k)) {
+                    result[k] = "[REDACTED]";
+                }
+                else {
+                    result[k] = ContentSanitizer.sanitize(v);
+                }
+            }
+            return result;
+        }
+        return value;
+    }
+}
+//# sourceMappingURL=content-sanitizer.js.map

package/dist/content/content-sanitizer.js.map

@@ -0,0 +1 @@
+{"version":3,"file":"content-sanitizer.js","sourceRoot":"","sources":["../../src/content/content-sanitizer.ts"],"names":[],"mappings":"AACA,OAAO,EAAE,mBAAmB,EAAE,MAAM,sBAAsB,CAAC;AAa3D,MAAM,OAAO,gBAAgB;IACnB,MAAM,CAAU,eAAe,GAIlC;QACH;YACE,IAAI,EAAE,aAAa;YACnB,KAAK,EAAE,6EAA6E;YACpF,SAAS,EAAE,wBAAwB;SACpC;QACD;YACE,IAAI,EAAE,gBAAgB;YACtB,KAAK,EAAE,6BAA6B;YACpC,SAAS,EAAE,2BAA2B;SACvC;QACD;YACE,IAAI,EAAE,YAAY;YAClB,KAAK,EAAE,+EAA+E;YACtF,SAAS,EAAE,uBAAuB;SACnC;QACD;YACE,IAAI,EAAE,gBAAgB;YACtB,KAAK,EAAE,uBAAuB;YAC9B,SAAS,EAAE,2BAA2B;SACvC;QACD;YACE,IAAI,EAAE,cAAc;YACpB,KAAK,EAAE,sCAAsC;YAC7C,SAAS,EAAE,gCAAgC;SAC5C;QACD;YACE,IAAI,EAAE,eAAe;YACrB,KAAK,EAAE,+EAA+E;YACtF,SAAS,EAAE,0BAA0B;SACtC;KACF,CAAC;IAEF;;;OAGG;IACI,MAAM,CAAC,WAAW,CAAC,IAAY;QACpC,IAAI,CAAC,IAAI;YAAE,OAAO,EAAE,UAAU,EAAE,KAAK,EAAE,QAAQ,EAAE,EAAE,EAAE,CAAC;QAEtD,MAAM,QAAQ,GAAoB,EAAE,CAAC;QAErC,KAAK,MAAM,EAAE,IAAI,EAAE,KAAK,EAAE,IAAI,gBAAgB,CAAC,eAAe,EAAE,CAAC;YAC/D,KAAK,CAAC,SAAS,GAAG,CAAC,CAAC;YACpB,IAAI,KAA6B,CAAC;YAClC,OAAO,CAAC,KAAK,GAAG,KAAK,CAAC,IAAI,CAAC,IAAI,CAAC,CAAC,KAAK,IAAI,EAAE,CAAC;gBAC3C,QAAQ,CAAC,IAAI,CAAC;oBACZ,IAAI;oBACJ,cAAc,EAAE,KAAK,CAAC,CAAC,CAAC,CAAC,KAAK,CAAC,CAAC,EAAE,CAAC,CAAC,GAAG,KAAK;oBAC5C,KAAK,EAAE,KAAK,CAAC,KAAK;iBACnB,CAAC,CAAC;YACL,CAAC;QACH,CAAC;QAED,OAAO;YACL,UAAU,EAAE,QAAQ,CAAC,MAAM,GAAG,CAAC;YAC/B,QAAQ;SACT,CAAC;IACJ,CAAC;IAED;;OAEG;IACI,MAAM,CAAC,aAAa,CAAC,IAAY;QACtC,IAAI,CAAC,IAAI;YAAE,OAAO,EAAE,YAAY,EAAE,EAAE,EAAE,QAAQ,EAAE,EAAE,EAAE,CAAC;QAErD,IAAI,YAAY,GAAG,IAAI,CAAC;QACxB,MAAM,QAAQ,GAAoB,EAAE,CAAC;QAErC,KAAK,MAAM,EAAE,IAAI,EAAE,KAAK,EAAE,SAAS,EAAE,IAAI,gBAAgB,CAAC,eAAe,EAAE,CAAC;YAC1E,KAAK,CAAC,SAAS,GAAG,CAAC,CAAC;YACpB,IAAI,KAA6B,CAAC;YAClC,OAAO,CAAC,KAAK,GAAG,KAAK,CAAC,IAAI,CAAC,IAAI,CAAC,CAAC,KAAK,IAAI,EAAE,CAAC;gBAC3C,QAAQ,CAAC,IAAI,CAAC;oBACZ,IAAI;oBACJ,cAAc,EAAE,KAAK,CAAC,CAAC,CAAC,CAAC,KAAK,CAAC,CAAC,EAAE,CAAC,CAAC,GAAG,KAAK;oBAC5C,KAAK,EAAE,KAAK,CAAC,KAAK;iBACnB,CAAC,CAAC;YACL,CAAC;YACD,YAAY,GAAG,YAAY,CAAC,OAAO,CAAC,KAAK,EAAE,SAAS,CAAC,CAAC;QACxD,CAAC;QAED,OAAO;YACL,YAAY;YACZ,QAAQ;SACT,CAAC;IACJ,CAAC;IAED;;;OAGG;IACI,MAAM,CAAC,qBAAqB,CAAC,OAAsB;QAKxD,IAAI,aAAa,GAAG,CAAC,CAAC;QACtB,MAAM,wBAAwB,GAA4B,EAAE,CAAC;QAE7D,KAAK,MAAM,GAAG,IAAI,OAAO,CAAC,eAAe,EAAE,CAAC;YAC1C,IAAI,GAAG,CAAC,IAAI,KAAK,MAAM,IAAI,GAAG,CAAC,IAAI,KAAK,UAAU,IAAI,GAAG,CAAC,IAAI,KAAK,YAAY,EAAE,CAAC;gBAChF,IAAI,OAAO,GAAG,CAAC,IAAI,KAAK,QAAQ,EAAE,CAAC;oBACjC,MAAM,EAAE,YAAY,EAAE,QAAQ,EAAE,GAAG,gBAAgB,CAAC,aAAa,CAAC,GAAG,CAAC,IAAI,CAAC,CAAC;oBAC5E,aAAa,IAAI,QAAQ,CAAC,MAAM,CAAC;oBACjC,wBAAwB,CAAC,IAAI,CAAC;wBAC5B,GAAG,GAAG;wBACN,IAAI,EAAE,YAAY;qBACnB,CAAC,CAAC;oBACH,SAAS;gBACX,CAAC;YACH,CAAC;YACD,wBAAwB,CAAC,IAAI,CAAC,GAAG,CAAC,CAAC;QACrC,CAAC;QAED,MAAM,YAAY,GAAG,aAAa,GAAG,CAAC,CAAC;QACvC,MAAM,WAAW,GAAG,YAAY;YAC9B,CAAC,CAAC,CAAC,OAAO,CAAC,QAAQ,CAAC,WAAW,KAAK,QAAQ,CAAC,CAAC,CAAC,QAAQ,CAAC,CAAC,CAAC,WAAW,CAAC;YACtE,CAAC,CAAC,OAAO,CAAC,QAAQ,CAAC,WAAW,CAAC;QAEjC,MAAM,SAAS,GAAkB;YAC/B,GAAG,OAAO;YACV,eAAe,EAAE,wBAAwB;YACzC,QAAQ,EAAE;gBACR,GAAG,OAAO,CAAC,QAAQ;gBACnB,WAAW;gBACX,OAAO,EAAE,IAAI;gBACb,WAAW,EAAE,OAAO;aACrB;YACD,SAAS,EAAE,IAAI,IAAI,EAAE,CAAC,WAAW,EAAE;SACpC,CAAC;QAEF,OAAO;YACL,SAAS,EAAE,MAAM,CAAC,MAAM,CAAC,mBAAmB,CAAC,KAAK,CAAC,SAAS,CAAC,CAAC;YAC9D,YAAY;YACZ,aAAa,EAAE,aAAa;SAC7B,CAAC;IACJ,CAAC;IAED;;OAEG;IACI,MAAM,CAAC,QAAQ,CAAC,KAAc;QACnC,IAAI,KAAK,KAAK,IAAI,IAAI,KAAK,KAAK,SAAS,EAAE,CAAC;YAC1C,OAAO,KAAK,CAAC;QACf,CAAC;QAED,IAAI,OAAO,KAAK,KAAK,QAAQ,EAAE,CAAC;YAC9B,OAAO,gBAAgB,CAAC,aAAa,CAAC,KAAK,CAAC,CAAC,YAAY,CAAC;QAC5D,CAAC;QAED,IAAI,KAAK,CAAC,OAAO,CAAC,KAAK,CAAC,EAAE,CAAC;YACzB,OAAO,KAAK,CAAC,GAAG,CAAC,CAAC,IAAI,EAAE,EAAE,CAAC,gBAAgB,CAAC,QAAQ,CAAC,IAAI,CAAC,CAAC,CAAC;QAC9D,CAAC;QAED,IAAI,OAAO,KAAK,KAAK,QAAQ,EAAE,CAAC;YAC9B,MAAM,MAAM,GAA4B,EAAE,CAAC;YAC3C,MAAM,mBAAmB,GAAG,6DAA6D,CAAC;YAE1F,KAAK,MAAM,CAAC,CAAC,EAAE,CAAC,CAAC,IAAI,MAAM,CAAC,OAAO,CAAC,KAAgC,CAAC,EAAE,CAAC;gBACtE,IAAI,mBAAmB,CAAC,IAAI,CAAC,CAAC,CAAC,EAAE,CAAC;oBAChC,MAAM,CAAC,CAAC,CAAC,GAAG,YAAY,CAAC;gBAC3B,CAAC;qBAAM,CAAC;oBACN,MAAM,CAAC,CAAC,CAAC,GAAG,gBAAgB,CAAC,QAAQ,CAAC,CAAC,CAAC,CAAC;gBAC3C,CAAC;YACH,CAAC;YACD,OAAO,MAAM,CAAC;QAChB,CAAC;QAGD,OAAO,KAAK,CAAC;IACf,CAAC"}

package/dist/content/index.d.ts

@@ -0,0 +1,15 @@
+export * from "./content-guards.js";
+export * from "./content-ingestion-engine.js";
+export * from "./representation-selector.js";
+export * from "./content-reference-manager.js";
+export * from "./content-access-validator.js";
+export * from "./content-sanitizer.js";
+export * from "./provenance-manager.js";
+export * from "./parsers/text-parser.js";
+export * from "./parsers/structured-parser.js";
+export * from "./parsers/pdf-parser.js";
+export * from "./parsers/image-parser.js";
+export * from "./parsers/media-parser.js";
+export * from "./parsers/archive-parser.js";
+export * from "./parsers/binary-parser.js";
+//# sourceMappingURL=index.d.ts.map

package/dist/content/index.d.ts.map

@@ -0,0 +1 @@
+{"version":3,"file":"index.d.ts","sourceRoot":"","sources":["../../src/content/index.ts"],"names":[],"mappings":"AAAA,cAAc,qBAAqB,CAAC;AACpC,cAAc,+BAA+B,CAAC;AAC9C,cAAc,8BAA8B,CAAC;AAC7C,cAAc,gCAAgC,CAAC;AAC/C,cAAc,+BAA+B,CAAC;AAC9C,cAAc,wBAAwB,CAAC;AACvC,cAAc,yBAAyB,CAAC;AACxC,cAAc,0BAA0B,CAAC;AACzC,cAAc,gCAAgC,CAAC;AAC/C,cAAc,yBAAyB,CAAC;AACxC,cAAc,2BAA2B,CAAC;AAC1C,cAAc,2BAA2B,CAAC;AAC1C,cAAc,6BAA6B,CAAC;AAC5C,cAAc,4BAA4B,CAAC"}

package/dist/content/index.js

@@ -0,0 +1,15 @@
+export * from "./content-guards.js";
+export * from "./content-ingestion-engine.js";
+export * from "./representation-selector.js";
+export * from "./content-reference-manager.js";
+export * from "./content-access-validator.js";
+export * from "./content-sanitizer.js";
+export * from "./provenance-manager.js";
+export * from "./parsers/text-parser.js";
+export * from "./parsers/structured-parser.js";
+export * from "./parsers/pdf-parser.js";
+export * from "./parsers/image-parser.js";
+export * from "./parsers/media-parser.js";
+export * from "./parsers/archive-parser.js";
+export * from "./parsers/binary-parser.js";
+//# sourceMappingURL=index.js.map

package/dist/content/index.js.map

@@ -0,0 +1 @@
+{"version":3,"file":"index.js","sourceRoot":"","sources":["../../src/content/index.ts"],"names":[],"mappings":"AAAA,cAAc,qBAAqB,CAAC;AACpC,cAAc,+BAA+B,CAAC;AAC9C,cAAc,8BAA8B,CAAC;AAC7C,cAAc,gCAAgC,CAAC;AAC/C,cAAc,+BAA+B,CAAC;AAC9C,cAAc,wBAAwB,CAAC;AACvC,cAAc,yBAAyB,CAAC;AACxC,cAAc,0BAA0B,CAAC;AACzC,cAAc,gCAAgC,CAAC;AAC/C,cAAc,yBAAyB,CAAC;AACxC,cAAc,2BAA2B,CAAC;AAC1C,cAAc,2BAA2B,CAAC;AAC1C,cAAc,6BAA6B,CAAC;AAC5C,cAAc,4BAA4B,CAAC"}

package/dist/content/parsers/archive-parser.d.ts

@@ -0,0 +1,25 @@
+import type { ContentRepresentation } from "../../domain/content.js";
+export interface ArchiveEntry {
+    name: string;
+    sizeBytes: number;
+    isDirectory: boolean;
+    compressedSizeBytes?: number;
+}
+export interface ArchiveParseResult {
+    representations: ContentRepresentation[];
+    metadata: {
+        format: "zip" | "tar" | "gzip";
+        entryCount: number;
+        totalUncompressedBytes: number;
+        isSafe: boolean;
+        violations: string[];
+    };
+}
+export declare class ArchiveParser {
+    /**
+     * Safely indexes archive contents and inspects entry tables without unsafe extraction.
+     * PRD Part 1 Section 13 & PRD Part 3 Section 138.
+     */
+    static parse(buffer: Buffer, mimeType: string): ArchiveParseResult;
+}
+//# sourceMappingURL=archive-parser.d.ts.map

package/dist/content/parsers/archive-parser.d.ts.map

@@ -0,0 +1 @@
+{"version":3,"file":"archive-parser.d.ts","sourceRoot":"","sources":["../../../src/content/parsers/archive-parser.ts"],"names":[],"mappings":"AACA,OAAO,KAAK,EAAE,qBAAqB,EAAE,MAAM,yBAAyB,CAAC;AAGrE,MAAM,WAAW,YAAY;IAC3B,IAAI,EAAE,MAAM,CAAC;IACb,SAAS,EAAE,MAAM,CAAC;IAClB,WAAW,EAAE,OAAO,CAAC;IACrB,mBAAmB,CAAC,EAAE,MAAM,CAAC;CAC9B;AAED,MAAM,WAAW,kBAAkB;IACjC,eAAe,EAAE,qBAAqB,EAAE,CAAC;IACzC,QAAQ,EAAE;QACR,MAAM,EAAE,KAAK,GAAG,KAAK,GAAG,MAAM,CAAC;QAC/B,UAAU,EAAE,MAAM,CAAC;QACnB,sBAAsB,EAAE,MAAM,CAAC;QAC/B,MAAM,EAAE,OAAO,CAAC;QAChB,UAAU,EAAE,MAAM,EAAE,CAAC;KACtB,CAAC;CACH;AAED,qBAAa,aAAa;IACxB;;;OAGG;WACW,KAAK,CAAC,MAAM,EAAE,MAAM,EAAE,QAAQ,EAAE,MAAM,GAAG,kBAAkB;CAyF1E"}

package/dist/content/parsers/archive-parser.js

@@ -0,0 +1,82 @@
+import { createHash } from "node:crypto";
+import { ContentGuards } from "../content-guards.js";
+export class ArchiveParser {
+    /**
+     * Safely indexes archive contents and inspects entry tables without unsafe extraction.
+     * PRD Part 1 Section 13 & PRD Part 3 Section 138.
+     */
+    static parse(buffer, mimeType) {
+        const sha256 = createHash("sha256").update(buffer).digest("hex");
+        const entries = [];
+        let format = "zip";
+        if (mimeType.includes("gzip") || (buffer.length >= 2 && buffer[0] === 0x1f && buffer[1] === 0x8b)) {
+            format = "gzip";
+        }
+        // Parse ZIP Local File Headers: PK\x03\x04
+        if (format === "zip") {
+            let offset = 0;
+            while (offset < buffer.length - 30) {
+                if (buffer[offset] === 0x50 &&
+                    buffer[offset + 1] === 0x4b &&
+                    buffer[offset + 2] === 0x03 &&
+                    buffer[offset + 3] === 0x04) {
+                    const compSize = buffer.readUInt32LE(offset + 18);
+                    const uncompSize = buffer.readUInt32LE(offset + 22);
+                    const nameLen = buffer.readUInt16LE(offset + 26);
+                    const extraLen = buffer.readUInt16LE(offset + 28);
+                    if (offset + 30 + nameLen <= buffer.length) {
+                        const entryName = buffer.toString("utf8", offset + 30, offset + 30 + nameLen);
+                        const isDirectory = entryName.endsWith("/");
+                        entries.push({
+                            name: entryName,
+                            sizeBytes: uncompSize,
+                            compressedSizeBytes: compSize,
+                            isDirectory,
+                        });
+                        // Advance offset past local header + name + extra + compressed payload
+                        offset += 30 + nameLen + extraLen + compSize;
+                        continue;
+                    }
+                }
+                offset++;
+            }
+        }
+        const entryNames = entries.map((e) => e.name);
+        const safety = ContentGuards.checkArchiveSafety(entryNames);
+        const totalUncompressedBytes = entries.reduce((acc, e) => acc + e.sizeBytes, 0);
+        const representations = [];
+        // Archive Index Representation
+        const indexData = JSON.stringify({
+            format,
+            entryCount: entries.length,
+            totalUncompressedBytes,
+            entries,
+            safety,
+        }, null, 2);
+        representations.push({
+            id: `rep_arc_${sha256.slice(0, 12)}`,
+            type: "archive-index",
+            mimeType: "application/json",
+            sizeBytes: Buffer.byteLength(indexData, "utf8"),
+            sha256: createHash("sha256").update(indexData).digest("hex"),
+            data: indexData,
+            metadata: {
+                format,
+                entryCount: entries.length,
+                totalUncompressedBytes,
+                isSafe: safety.isSafe,
+            },
+        });
+        return {
+            representations,
+            metadata: {
+                format,
+                entryCount: entries.length,
+                totalUncompressedBytes,
+                isSafe: safety.isSafe,
+                violations: safety.violations,
+            },
+        };
+    }
+}
+//# sourceMappingURL=archive-parser.js.map

package/dist/content/parsers/archive-parser.js.map

@@ -0,0 +1 @@
+{"version":3,"file":"archive-parser.js","sourceRoot":"","sources":["../../../src/content/parsers/archive-parser.ts"],"names":[],"mappings":"AAAA,OAAO,EAAE,UAAU,EAAE,MAAM,aAAa,CAAC;AAEzC,OAAO,EAAE,aAAa,EAAE,MAAM,sBAAsB,CAAC;AAoBrD,MAAM,OAAO,aAAa;IACxB;;;OAGG;IACI,MAAM,CAAC,KAAK,CAAC,MAAc,EAAE,QAAgB;QAClD,MAAM,MAAM,GAAG,UAAU,CAAC,QAAQ,CAAC,CAAC,MAAM,CAAC,MAAM,CAAC,CAAC,MAAM,CAAC,KAAK,CAAC,CAAC;QACjE,MAAM,OAAO,GAAmB,EAAE,CAAC;QACnC,IAAI,MAAM,GAA2B,KAAK,CAAC;QAE3C,IAAI,QAAQ,CAAC,QAAQ,CAAC,MAAM,CAAC,IAAI,CAAC,MAAM,CAAC,MAAM,IAAI,CAAC,IAAI,MAAM,CAAC,CAAC,CAAC,KAAK,IAAI,IAAI,MAAM,CAAC,CAAC,CAAC,KAAK,IAAI,CAAC,EAAE,CAAC;YAClG,MAAM,GAAG,MAAM,CAAC;QAClB,CAAC;QAED,2CAA2C;QAC3C,IAAI,MAAM,KAAK,KAAK,EAAE,CAAC;YACrB,IAAI,MAAM,GAAG,CAAC,CAAC;YACf,OAAO,MAAM,GAAG,MAAM,CAAC,MAAM,GAAG,EAAE,EAAE,CAAC;gBACnC,IACE,MAAM,CAAC,MAAM,CAAC,KAAK,IAAI;oBACvB,MAAM,CAAC,MAAM,GAAG,CAAC,CAAC,KAAK,IAAI;oBAC3B,MAAM,CAAC,MAAM,GAAG,CAAC,CAAC,KAAK,IAAI;oBAC3B,MAAM,CAAC,MAAM,GAAG,CAAC,CAAC,KAAK,IAAI,EAC3B,CAAC;oBACD,MAAM,QAAQ,GAAG,MAAM,CAAC,YAAY,CAAC,MAAM,GAAG,EAAE,CAAC,CAAC;oBAClD,MAAM,UAAU,GAAG,MAAM,CAAC,YAAY,CAAC,MAAM,GAAG,EAAE,CAAC,CAAC;oBACpD,MAAM,OAAO,GAAG,MAAM,CAAC,YAAY,CAAC,MAAM,GAAG,EAAE,CAAC,CAAC;oBACjD,MAAM,QAAQ,GAAG,MAAM,CAAC,YAAY,CAAC,MAAM,GAAG,EAAE,CAAC,CAAC;oBAElD,IAAI,MAAM,GAAG,EAAE,GAAG,OAAO,IAAI,MAAM,CAAC,MAAM,EAAE,CAAC;wBAC3C,MAAM,SAAS,GAAG,MAAM,CAAC,QAAQ,CAAC,MAAM,EAAE,MAAM,GAAG,EAAE,EAAE,MAAM,GAAG,EAAE,GAAG,OAAO,CAAC,CAAC;wBAC9E,MAAM,WAAW,GAAG,SAAS,CAAC,QAAQ,CAAC,GAAG,CAAC,CAAC;wBAE5C,OAAO,CAAC,IAAI,CAAC;4BACX,IAAI,EAAE,SAAS;4BACf,SAAS,EAAE,UAAU;4BACrB,mBAAmB,EAAE,QAAQ;4BAC7B,WAAW;yBACZ,CAAC,CAAC;wBAEH,uEAAuE;wBACvE,MAAM,IAAI,EAAE,GAAG,OAAO,GAAG,QAAQ,GAAG,QAAQ,CAAC;wBAC7C,SAAS;oBACX,CAAC;gBACH,CAAC;gBACD,MAAM,EAAE,CAAC;YACX,CAAC;QACH,CAAC;QAED,MAAM,UAAU,GAAG,OAAO,CAAC,GAAG,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC,IAAI,CAAC,CAAC;QAC9C,MAAM,MAAM,GAAG,aAAa,CAAC,kBAAkB,CAAC,UAAU,CAAC,CAAC;QAC5D,MAAM,sBAAsB,GAAG,OAAO,CAAC,MAAM,CAAC,CAAC,GAAG,EAAE,CAAC,EAAE,EAAE,CAAC,GAAG,GAAG,CAAC,CAAC,SAAS,EAAE,CAAC,CAAC,CAAC;QAEhF,MAAM,eAAe,GAA4B,EAAE,CAAC;QAEpD,+BAA+B;QAC/B,MAAM,SAAS,GAAG,IAAI,CAAC,SAAS,CAC9B;YACE,MAAM;YACN,UAAU,EAAE,OAAO,CAAC,MAAM;YAC1B,sBAAsB;YACtB,OAAO;YACP,MAAM;SACP,EACD,IAAI,EACJ,CAAC,CACF,CAAC;QAEF,eAAe,CAAC,IAAI,CAAC;YACnB,EAAE,EAAE,WAAW,MAAM,CAAC,KAAK,CAAC,CAAC,EAAE,EAAE,CAAC,EAAE;YACpC,IAAI,EAAE,eAAe;YACrB,QAAQ,EAAE,kBAAkB;YAC5B,SAAS,EAAE,MAAM,CAAC,UAAU,CAAC,SAAS,EAAE,MAAM,CAAC;YAC/C,MAAM,EAAE,UAAU,CAAC,QAAQ,CAAC,CAAC,MAAM,CAAC,SAAS,CAAC,CAAC,MAAM,CAAC,KAAK,CAAC;YAC5D,IAAI,EAAE,SAAS;YACf,QAAQ,EAAE;gBACR,MAAM;gBACN,UAAU,EAAE,OAAO,CAAC,MAAM;gBAC1B,sBAAsB;gBACtB,MAAM,EAAE,MAAM,CAAC,MAAM;aACtB;SACF,CAAC,CAAC;QAEH,OAAO;YACL,eAAe;YACf,QAAQ,EAAE;gBACR,MAAM;gBACN,UAAU,EAAE,OAAO,CAAC,MAAM;gBAC1B,sBAAsB;gBACtB,MAAM,EAAE,MAAM,CAAC,MAAM;gBACrB,UAAU,EAAE,MAAM,CAAC,UAAU;aAC9B;SACF,CAAC;IACJ,CAAC;CACF"}