@monoes/monomindcli 2.14.0 → 2.15.0

package/dist/src/knowledge/mhtml.js

@@ -0,0 +1,154 @@
+/**
+ * MHTML / MHT container parsing (RCL-01).
+ *
+ * An MHTML archive is an RFC 2557 `multipart/related` MIME message: the page
+ * plus every subresource it referenced, each part carrying its own transfer
+ * encoding and charset. Chrome's `Page.captureSnapshot` writes them
+ * quoted-printable; Internet Explorer and most "save as web archive" tools
+ * write base64.
+ *
+ * Only the ROOT frame is prose. Picking the first `text/html` part instead is
+ * how an ad iframe ends up indexed as the article, so the root is resolved
+ * from the container's `Snapshot-Content-Location` / `Content-Location` when
+ * one is present.
+ *
+ * Read the file as `latin1` (or pass a Buffer): byte fidelity matters before
+ * transfer decoding, and the real charset is only known per part.
+ *
+ * @module v1/cli/knowledge/mhtml
+ */
+/** Decode a quoted-printable body to bytes: soft line breaks vanish, `=XX`
+ *  becomes one byte, everything else is its own latin1 byte. */
+export function decodeQuotedPrintable(input) {
+    const out = [];
+    for (let i = 0; i < input.length; i++) {
+        const ch = input[i];
+        if (ch !== '=') {
+            out.push(input.charCodeAt(i) & 0xff);
+            continue;
+        }
+        if (input[i + 1] === '\r' && input[i + 2] === '\n') {
+            i += 2; // soft line break
+            continue;
+        }
+        if (input[i + 1] === '\n') {
+            i += 1; // soft line break, LF-only file
+            continue;
+        }
+        const hex = input.slice(i + 1, i + 3);
+        if (/^[0-9a-fA-F]{2}$/.test(hex)) {
+            out.push(Number.parseInt(hex, 16));
+            i += 2;
+            continue;
+        }
+        out.push(0x3d); // a lone `=` — keep it
+    }
+    return Buffer.from(out);
+}
+function decodeBody(raw, encoding) {
+    switch (encoding) {
+        case 'quoted-printable':
+            return decodeQuotedPrintable(raw);
+        case 'base64':
+            return Buffer.from(raw.replace(/\s+/g, ''), 'base64');
+        default:
+            return Buffer.from(raw, 'latin1');
+    }
+}
+/** Decode bytes under a MIME charset, falling back to UTF-8 (then latin1)
+ *  rather than failing an ingest over an exotic label. */
+export function decodeCharset(bytes, charset) {
+    const label = (charset ?? 'utf-8').trim().toLowerCase();
+    for (const candidate of [label, 'utf-8']) {
+        try {
+            return new TextDecoder(candidate, { fatal: false }).decode(bytes);
+        }
+        catch {
+            /* unknown label — try the next */
+        }
+    }
+    return bytes.toString('latin1');
+}
+/** Split a MIME chunk into unfolded headers and the body after the first
+ *  blank line. Continuation lines (leading space/tab) join their header. */
+function splitHeaders(chunk) {
+    const sep = /\r?\n\r?\n/.exec(chunk);
+    const headerText = sep ? chunk.slice(0, sep.index) : chunk;
+    const body = sep ? chunk.slice(sep.index + sep[0].length) : '';
+    const headers = {};
+    let current = '';
+    for (const rawLine of headerText.split(/\r?\n/)) {
+        if (/^[ \t]/.test(rawLine) && current) {
+            headers[current] += ` ${rawLine.trim()}`;
+            continue;
+        }
+        const colon = rawLine.indexOf(':');
+        if (colon === -1)
+            continue;
+        current = rawLine.slice(0, colon).trim().toLowerCase();
+        headers[current] = rawLine.slice(colon + 1).trim();
+    }
+    return { headers, body };
+}
+function param(headerValue, name) {
+    if (!headerValue)
+        return null;
+    const m = new RegExp(`${name}\\s*=\\s*("[^"]*"|'[^']*'|[^;\\s]+)`, 'i').exec(headerValue);
+    return m ? m[1].replace(/^["']|["']$/g, '') : null;
+}
+function toPart(chunk) {
+    const { headers, body } = splitHeaders(chunk);
+    const contentType = (headers['content-type'] ?? 'text/html').split(';')[0].trim().toLowerCase();
+    const encoding = (headers['content-transfer-encoding'] ?? 'binary').trim().toLowerCase();
+    return {
+        contentType,
+        charset: param(headers['content-type'], 'charset')?.toLowerCase() ?? null,
+        location: headers['content-location'] ?? null,
+        encoding,
+        body: decodeBody(body, encoding),
+        headers,
+    };
+}
+/**
+ * Parse an MHTML container. Anything that is not a well-formed multipart
+ * message is treated as a single part, so a plain `.html` file renamed to
+ * `.mhtml` still ingests.
+ */
+export function parseMhtml(raw) {
+    const text = typeof raw === 'string' ? raw : raw.toString('latin1');
+    const { headers, body } = splitHeaders(text);
+    const snapshotLocation = headers['snapshot-content-location'] ?? headers['content-location'] ?? null;
+    const subject = headers.subject ?? null;
+    const boundary = param(headers['content-type'], 'boundary');
+    let parts;
+    if (!boundary) {
+        // No boundary: either a single-part message or a bare HTML file. Feeding
+        // the whole document through `toPart` covers both — an HTML file has no
+        // header block, so everything lands in the body.
+        parts = [toPart(/^[a-z-]+:/i.test(text) ? text : `Content-Type: text/html\r\n\r\n${text}`)];
+    }
+    else {
+        const delimiter = `--${boundary}`;
+        parts = body
+            .split(delimiter)
+            .slice(1) // text before the first delimiter is the MIME preamble
+            .filter((chunk) => !chunk.startsWith('--')) // the closing `--boundary--`
+            .map((chunk) => chunk.replace(/^\r?\n/, ''))
+            .map(toPart);
+    }
+    const html = parts.filter((p) => p.contentType === 'text/html' || p.contentType === 'application/xhtml+xml');
+    const root = (snapshotLocation ? html.find((p) => p.location === snapshotLocation) : undefined) ??
+        html[0] ??
+        null;
+    return { root, parts, snapshotLocation, subject };
+}
+/** Parse a container and return its root frame decoded to a string. */
+export function mhtmlToHtml(raw) {
+    const parsed = parseMhtml(raw);
+    return {
+        html: parsed.root ? decodeCharset(parsed.root.body, parsed.root.charset) : '',
+        location: parsed.root?.location ?? parsed.snapshotLocation,
+        subject: parsed.subject,
+    };
+}
+//# sourceMappingURL=mhtml.js.map

package/dist/src/knowledge/mhtml.js.map

@@ -0,0 +1 @@
+{"version":3,"file":"mhtml.js","sourceRoot":"","sources":["../../../src/knowledge/mhtml.ts"],"names":[],"mappings":"AAAA;;;;;;;;;;;;;;;;;;GAkBG;AA0BH;gEACgE;AAChE,MAAM,UAAU,qBAAqB,CAAC,KAAa;IACjD,MAAM,GAAG,GAAa,EAAE,CAAC;IACzB,KAAK,IAAI,CAAC,GAAG,CAAC,EAAE,CAAC,GAAG,KAAK,CAAC,MAAM,EAAE,CAAC,EAAE,EAAE,CAAC;QACtC,MAAM,EAAE,GAAG,KAAK,CAAC,CAAC,CAAC,CAAC;QACpB,IAAI,EAAE,KAAK,GAAG,EAAE,CAAC;YACf,GAAG,CAAC,IAAI,CAAC,KAAK,CAAC,UAAU,CAAC,CAAC,CAAC,GAAG,IAAI,CAAC,CAAC;YACrC,SAAS;QACX,CAAC;QACD,IAAI,KAAK,CAAC,CAAC,GAAG,CAAC,CAAC,KAAK,IAAI,IAAI,KAAK,CAAC,CAAC,GAAG,CAAC,CAAC,KAAK,IAAI,EAAE,CAAC;YACnD,CAAC,IAAI,CAAC,CAAC,CAAC,kBAAkB;YAC1B,SAAS;QACX,CAAC;QACD,IAAI,KAAK,CAAC,CAAC,GAAG,CAAC,CAAC,KAAK,IAAI,EAAE,CAAC;YAC1B,CAAC,IAAI,CAAC,CAAC,CAAC,gCAAgC;YACxC,SAAS;QACX,CAAC;QACD,MAAM,GAAG,GAAG,KAAK,CAAC,KAAK,CAAC,CAAC,GAAG,CAAC,EAAE,CAAC,GAAG,CAAC,CAAC,CAAC;QACtC,IAAI,kBAAkB,CAAC,IAAI,CAAC,GAAG,CAAC,EAAE,CAAC;YACjC,GAAG,CAAC,IAAI,CAAC,MAAM,CAAC,QAAQ,CAAC,GAAG,EAAE,EAAE,CAAC,CAAC,CAAC;YACnC,CAAC,IAAI,CAAC,CAAC;YACP,SAAS;QACX,CAAC;QACD,GAAG,CAAC,IAAI,CAAC,IAAI,CAAC,CAAC,CAAC,uBAAuB;IACzC,CAAC;IACD,OAAO,MAAM,CAAC,IAAI,CAAC,GAAG,CAAC,CAAC;AAC1B,CAAC;AAED,SAAS,UAAU,CAAC,GAAW,EAAE,QAAgB;IAC/C,QAAQ,QAAQ,EAAE,CAAC;QACjB,KAAK,kBAAkB;YACrB,OAAO,qBAAqB,CAAC,GAAG,CAAC,CAAC;QACpC,KAAK,QAAQ;YACX,OAAO,MAAM,CAAC,IAAI,CAAC,GAAG,CAAC,OAAO,CAAC,MAAM,EAAE,EAAE,CAAC,EAAE,QAAQ,CAAC,CAAC;QACxD;YACE,OAAO,MAAM,CAAC,IAAI,CAAC,GAAG,EAAE,QAAQ,CAAC,CAAC;IACtC,CAAC;AACH,CAAC;AAED;0DAC0D;AAC1D,MAAM,UAAU,aAAa,CAAC,KAAa,EAAE,OAAsB;IACjE,MAAM,KAAK,GAAG,CAAC,OAAO,IAAI,OAAO,CAAC,CAAC,IAAI,EAAE,CAAC,WAAW,EAAE,CAAC;IACxD,KAAK,MAAM,SAAS,IAAI,CAAC,KAAK,EAAE,OAAO,CAAC,EAAE,CAAC;QACzC,IAAI,CAAC;YACH,OAAO,IAAI,WAAW,CAAC,SAAS,EAAE,EAAE,KAAK,EAAE,KAAK,EAAE,CAAC,CAAC,MAAM,CAAC,KAAK,CAAC,CAAC;QACpE,CAAC;QAAC,MAAM,CAAC;YACP,kCAAkC;QACpC,CAAC;IACH,CAAC;IACD,OAAO,KAAK,CAAC,QAAQ,CAAC,QAAQ,CAAC,CAAC;AAClC,CAAC;AAED;4EAC4E;AAC5E,SAAS,YAAY,CAAC,KAAa;IACjC,MAAM,GAAG,GAAG,YAAY,CAAC,IAAI,CAAC,KAAK,CAAC,CAAC;IACrC,MAAM,UAAU,GAAG,GAAG,CAAC,CAAC,CAAC,KAAK,CAAC,KAAK,CAAC,CAAC,EAAE,GAAG,CAAC,KAAK,CAAC,CAAC,CAAC,CAAC,KAAK,CAAC;IAC3D,MAAM,IAAI,GAAG,GAAG,CAAC,CAAC,CAAC,KAAK,CAAC,KAAK,CAAC,GAAG,CAAC,KAAK,GAAG,GAAG,CAAC,CAAC,CAAC,CAAC,MAAM,CAAC,CAAC,CAAC,CAAC,EAAE,CAAC;IAE/D,MAAM,OAAO,GAA2B,EAAE,CAAC;IAC3C,IAAI,OAAO,GAAG,EAAE,CAAC;IACjB,KAAK,MAAM,OAAO,IAAI,UAAU,CAAC,KAAK,CAAC,OAAO,CAAC,EAAE,CAAC;QAChD,IAAI,QAAQ,CAAC,IAAI,CAAC,OAAO,CAAC,IAAI,OAAO,EAAE,CAAC;YACtC,OAAO,CAAC,OAAO,CAAC,IAAI,IAAI,OAAO,CAAC,IAAI,EAAE,EAAE,CAAC;YACzC,SAAS;QACX,CAAC;QACD,MAAM,KAAK,GAAG,OAAO,CAAC,OAAO,CAAC,GAAG,CAAC,CAAC;QACnC,IAAI,KAAK,KAAK,CAAC,CAAC;YAAE,SAAS;QAC3B,OAAO,GAAG,OAAO,CAAC,KAAK,CAAC,CAAC,EAAE,KAAK,CAAC,CAAC,IAAI,EAAE,CAAC,WAAW,EAAE,CAAC;QACvD,OAAO,CAAC,OAAO,CAAC,GAAG,OAAO,CAAC,KAAK,CAAC,KAAK,GAAG,CAAC,CAAC,CAAC,IAAI,EAAE,CAAC;IACrD,CAAC;IACD,OAAO,EAAE,OAAO,EAAE,IAAI,EAAE,CAAC;AAC3B,CAAC;AAED,SAAS,KAAK,CAAC,WAA+B,EAAE,IAAY;IAC1D,IAAI,CAAC,WAAW;QAAE,OAAO,IAAI,CAAC;IAC9B,MAAM,CAAC,GAAG,IAAI,MAAM,CAAC,GAAG,IAAI,qCAAqC,EAAE,GAAG,CAAC,CAAC,IAAI,CAAC,WAAW,CAAC,CAAC;IAC1F,OAAO,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,OAAO,CAAC,cAAc,EAAE,EAAE,CAAC,CAAC,CAAC,CAAC,IAAI,CAAC;AACrD,CAAC;AAED,SAAS,MAAM,CAAC,KAAa;IAC3B,MAAM,EAAE,OAAO,EAAE,IAAI,EAAE,GAAG,YAAY,CAAC,KAAK,CAAC,CAAC;IAC9C,MAAM,WAAW,GAAG,CAAC,OAAO,CAAC,cAAc,CAAC,IAAI,WAAW,CAAC,CAAC,KAAK,CAAC,GAAG,CAAC,CAAC,CAAC,CAAC,CAAC,IAAI,EAAE,CAAC,WAAW,EAAE,CAAC;IAChG,MAAM,QAAQ,GAAG,CAAC,OAAO,CAAC,2BAA2B,CAAC,IAAI,QAAQ,CAAC,CAAC,IAAI,EAAE,CAAC,WAAW,EAAE,CAAC;IACzF,OAAO;QACL,WAAW;QACX,OAAO,EAAE,KAAK,CAAC,OAAO,CAAC,cAAc,CAAC,EAAE,SAAS,CAAC,EAAE,WAAW,EAAE,IAAI,IAAI;QACzE,QAAQ,EAAE,OAAO,CAAC,kBAAkB,CAAC,IAAI,IAAI;QAC7C,QAAQ;QACR,IAAI,EAAE,UAAU,CAAC,IAAI,EAAE,QAAQ,CAAC;QAChC,OAAO;KACR,CAAC;AACJ,CAAC;AAED;;;;GAIG;AACH,MAAM,UAAU,UAAU,CAAC,GAAoB;IAC7C,MAAM,IAAI,GAAG,OAAO,GAAG,KAAK,QAAQ,CAAC,CAAC,CAAC,GAAG,CAAC,CAAC,CAAC,GAAG,CAAC,QAAQ,CAAC,QAAQ,CAAC,CAAC;IACpE,MAAM,EAAE,OAAO,EAAE,IAAI,EAAE,GAAG,YAAY,CAAC,IAAI,CAAC,CAAC;IAC7C,MAAM,gBAAgB,GACpB,OAAO,CAAC,2BAA2B,CAAC,IAAI,OAAO,CAAC,kBAAkB,CAAC,IAAI,IAAI,CAAC;IAC9E,MAAM,OAAO,GAAG,OAAO,CAAC,OAAO,IAAI,IAAI,CAAC;IACxC,MAAM,QAAQ,GAAG,KAAK,CAAC,OAAO,CAAC,cAAc,CAAC,EAAE,UAAU,CAAC,CAAC;IAE5D,IAAI,KAAkB,CAAC;IACvB,IAAI,CAAC,QAAQ,EAAE,CAAC;QACd,yEAAyE;QACzE,wEAAwE;QACxE,iDAAiD;QACjD,KAAK,GAAG,CAAC,MAAM,CAAC,YAAY,CAAC,IAAI,CAAC,IAAI,CAAC,CAAC,CAAC,CAAC,IAAI,CAAC,CAAC,CAAC,kCAAkC,IAAI,EAAE,CAAC,CAAC,CAAC;IAC9F,CAAC;SAAM,CAAC;QACN,MAAM,SAAS,GAAG,KAAK,QAAQ,EAAE,CAAC;QAClC,KAAK,GAAG,IAAI;aACT,KAAK,CAAC,SAAS,CAAC;aAChB,KAAK,CAAC,CAAC,CAAC,CAAC,uDAAuD;aAChE,MAAM,CAAC,CAAC,KAAK,EAAE,EAAE,CAAC,CAAC,KAAK,CAAC,UAAU,CAAC,IAAI,CAAC,CAAC,CAAC,6BAA6B;aACxE,GAAG,CAAC,CAAC,KAAK,EAAE,EAAE,CAAC,KAAK,CAAC,OAAO,CAAC,QAAQ,EAAE,EAAE,CAAC,CAAC;aAC3C,GAAG,CAAC,MAAM,CAAC,CAAC;IACjB,CAAC;IAED,MAAM,IAAI,GAAG,KAAK,CAAC,MAAM,CACvB,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC,WAAW,KAAK,WAAW,IAAI,CAAC,CAAC,WAAW,KAAK,uBAAuB,CAClF,CAAC;IACF,MAAM,IAAI,GACR,CAAC,gBAAgB,CAAC,CAAC,CAAC,IAAI,CAAC,IAAI,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC,QAAQ,KAAK,gBAAgB,CAAC,CAAC,CAAC,CAAC,SAAS,CAAC;QAClF,IAAI,CAAC,CAAC,CAAC;QACP,IAAI,CAAC;IAEP,OAAO,EAAE,IAAI,EAAE,KAAK,EAAE,gBAAgB,EAAE,OAAO,EAAE,CAAC;AACpD,CAAC;AAED,uEAAuE;AACvE,MAAM,UAAU,WAAW,CAAC,GAAoB;IAK9C,MAAM,MAAM,GAAG,UAAU,CAAC,GAAG,CAAC,CAAC;IAC/B,OAAO;QACL,IAAI,EAAE,MAAM,CAAC,IAAI,CAAC,CAAC,CAAC,aAAa,CAAC,MAAM,CAAC,IAAI,CAAC,IAAI,EAAE,MAAM,CAAC,IAAI,CAAC,OAAO,CAAC,CAAC,CAAC,CAAC,EAAE;QAC7E,QAAQ,EAAE,MAAM,CAAC,IAAI,EAAE,QAAQ,IAAI,MAAM,CAAC,gBAAgB;QAC1D,OAAO,EAAE,MAAM,CAAC,OAAO;KACxB,CAAC;AACJ,CAAC"}

package/dist/src/knowledge/okf-bundle.d.ts

@@ -0,0 +1,19 @@
+/**
+ * OKF bundles — export the indexed documents to a portable Markdown bundle
+ * (frontmatter + `index.md`), and ingest one back.
+ *
+ * A bundle is plain files, so a round trip goes back through the ordinary
+ * ingest path rather than restoring index records directly: whatever comes
+ * back is re-extracted, re-chunked and re-versioned like any other document.
+ *
+ * Split out of document-pipeline.ts.
+ *
+ * @module v1/cli/knowledge/okf-bundle
+ */
+import type { BatchIngestResult } from './document-types.js';
+export declare function exportToOKF(outputDir: string, rootDir?: string, scope?: string): Promise<{
+    exported: number;
+    outputDir: string;
+}>;
+export declare function importFromOKF(bundleDir: string, scope?: string, rootDir?: string): Promise<BatchIngestResult>;
+//# sourceMappingURL=okf-bundle.d.ts.map

package/dist/src/knowledge/okf-bundle.d.ts.map

@@ -0,0 +1 @@
+{"version":3,"file":"okf-bundle.d.ts","sourceRoot":"","sources":["../../../src/knowledge/okf-bundle.ts"],"names":[],"mappings":"AAAA;;;;;;;;;;;GAWG;AAYH,OAAO,KAAK,EAAE,iBAAiB,EAAE,MAAM,qBAAqB,CAAC;AAE7D,wBAAsB,WAAW,CAC/B,SAAS,EAAE,MAAM,EACjB,OAAO,SAAmB,EAC1B,KAAK,SAAW,GACf,OAAO,CAAC;IAAE,QAAQ,EAAE,MAAM,CAAC;IAAC,SAAS,EAAE,MAAM,CAAA;CAAE,CAAC,CA8DlD;AAID,wBAAsB,aAAa,CACjC,SAAS,EAAE,MAAM,EACjB,KAAK,SAAW,EAChB,OAAO,SAAmB,GACzB,OAAO,CAAC,iBAAiB,CAAC,CA4B5B"}

package/dist/src/knowledge/okf-bundle.js

@@ -0,0 +1,106 @@
+/**
+ * OKF bundles — export the indexed documents to a portable Markdown bundle
+ * (frontmatter + `index.md`), and ingest one back.
+ *
+ * A bundle is plain files, so a round trip goes back through the ordinary
+ * ingest path rather than restoring index records directly: whatever comes
+ * back is re-extracted, re-chunked and re-versioned like any other document.
+ *
+ * Split out of document-pipeline.ts.
+ *
+ * @module v1/cli/knowledge/okf-bundle
+ */
+import * as fs from 'node:fs';
+import * as path from 'node:path';
+import { extractText } from '../capabilities/cap-documents.js';
+// Static import is safe and deliberate: memory-bridge imports only node builtins
+// at module scope (everything heavy is lazy), and the project-root rule must not
+// be duplicated — two copies of "which directory is this project" is exactly the
+// bug this default exists to fix.
+import { getProjectRoot } from '../memory/memory-bridge.js';
+import { listDocuments } from './document-index.js';
+import { ingestDocument, toFileEntry } from './document-ingest.js';
+export async function exportToOKF(outputDir, rootDir = getProjectRoot(), scope = 'shared') {
+    const docs = listDocuments(rootDir, scope);
+    fs.mkdirSync(outputDir, { recursive: true });
+    let exported = 0;
+    const indexEntries = [];
+    for (const doc of docs) {
+        // Read original content
+        let content = '';
+        try {
+            if (fs.existsSync(doc.filePath)) {
+                const entry = toFileEntry(doc.filePath);
+                content = await extractText(entry);
+            }
+        }
+        catch {
+            continue;
+        }
+        if (!content)
+            continue;
+        const title = path.basename(doc.filePath, path.extname(doc.filePath));
+        const ext = path.extname(doc.filePath).toLowerCase();
+        const relativePath = path.relative(rootDir, doc.filePath);
+        const slug = title.replace(/[^a-zA-Z0-9._-]+/g, '-').toLowerCase();
+        const outFile = path.join(outputDir, `${slug}.md`);
+        const yamlEscape = (s) => /[:"'[\]{}#&*!|>%@`]/.test(s) ? `"${s.replace(/\\/g, '\\\\').replace(/"/g, '\\"')}"` : s;
+        const frontmatter = [
+            '---',
+            `type: Document`,
+            `title: ${yamlEscape(title)}`,
+            `description: ${yamlEscape(`Extracted from ${path.basename(doc.filePath)}`)}`,
+            `resource: ${yamlEscape(relativePath)}`,
+            `tags: ["document", ${yamlEscape(ext.slice(1))}]`,
+            `timestamp: ${yamlEscape(doc.indexedAt)}`,
+            `contentHash: ${yamlEscape(doc.contentHash)}`,
+            `chunkCount: ${doc.chunkCount}`,
+            '---',
+            '',
+        ].join('\n');
+        fs.writeFileSync(outFile, frontmatter + content, 'utf-8');
+        indexEntries.push(`* [${title}](${slug}.md) - ${path.basename(doc.filePath)} (${doc.chunkCount} chunks)`);
+        exported++;
+    }
+    // Write index.md
+    const indexContent = [
+        `# Knowledge Bundle`,
+        '',
+        `Exported from monomind on ${new Date().toISOString().slice(0, 10)}`,
+        '',
+        ...indexEntries,
+        '',
+    ].join('\n');
+    fs.writeFileSync(path.join(outputDir, 'index.md'), indexContent, 'utf-8');
+    return { exported, outputDir };
+}
+// ── OKF Import ─────────────────────────────────────────────────────
+export async function importFromOKF(bundleDir, scope = 'shared', rootDir = getProjectRoot()) {
+    const resolved = path.resolve(bundleDir);
+    const files = fs
+        .readdirSync(resolved)
+        .filter((f) => f.endsWith('.md') && f !== 'index.md' && f !== 'log.md')
+        .map((f) => path.join(resolved, f));
+    const result = {
+        filesProcessed: 0,
+        filesSkipped: 0,
+        totalChunks: 0,
+        errors: [],
+        results: [],
+    };
+    for (const file of files) {
+        const r = await ingestDocument(file, scope, rootDir);
+        result.results.push(r);
+        if (r.skipped) {
+            result.filesSkipped++;
+        }
+        else {
+            result.filesProcessed++;
+            result.totalChunks += r.chunksIndexed;
+        }
+        if (r.error && !r.skipped)
+            result.errors.push(`${r.filePath}: ${r.error}`);
+    }
+    return result;
+}
+//# sourceMappingURL=okf-bundle.js.map

package/dist/src/knowledge/okf-bundle.js.map

@@ -0,0 +1 @@
+{"version":3,"file":"okf-bundle.js","sourceRoot":"","sources":["../../../src/knowledge/okf-bundle.ts"],"names":[],"mappings":"AAAA;;;;;;;;;;;GAWG;AAEH,OAAO,KAAK,EAAE,MAAM,SAAS,CAAC;AAC9B,OAAO,KAAK,IAAI,MAAM,WAAW,CAAC;AAClC,OAAO,EAAE,WAAW,EAAE,MAAM,kCAAkC,CAAC;AAC/D,iFAAiF;AACjF,iFAAiF;AACjF,iFAAiF;AACjF,kCAAkC;AAClC,OAAO,EAAE,cAAc,EAAE,MAAM,4BAA4B,CAAC;AAC5D,OAAO,EAAE,aAAa,EAAE,MAAM,qBAAqB,CAAC;AACpD,OAAO,EAAE,cAAc,EAAE,WAAW,EAAE,MAAM,sBAAsB,CAAC;AAGnE,MAAM,CAAC,KAAK,UAAU,WAAW,CAC/B,SAAiB,EACjB,OAAO,GAAG,cAAc,EAAE,EAC1B,KAAK,GAAG,QAAQ;IAEhB,MAAM,IAAI,GAAG,aAAa,CAAC,OAAO,EAAE,KAAK,CAAC,CAAC;IAC3C,EAAE,CAAC,SAAS,CAAC,SAAS,EAAE,EAAE,SAAS,EAAE,IAAI,EAAE,CAAC,CAAC;IAE7C,IAAI,QAAQ,GAAG,CAAC,CAAC;IACjB,MAAM,YAAY,GAAa,EAAE,CAAC;IAElC,KAAK,MAAM,GAAG,IAAI,IAAI,EAAE,CAAC;QACvB,wBAAwB;QACxB,IAAI,OAAO,GAAG,EAAE,CAAC;QACjB,IAAI,CAAC;YACH,IAAI,EAAE,CAAC,UAAU,CAAC,GAAG,CAAC,QAAQ,CAAC,EAAE,CAAC;gBAChC,MAAM,KAAK,GAAG,WAAW,CAAC,GAAG,CAAC,QAAQ,CAAC,CAAC;gBACxC,OAAO,GAAG,MAAM,WAAW,CAAC,KAAK,CAAC,CAAC;YACrC,CAAC;QACH,CAAC;QAAC,MAAM,CAAC;YACP,SAAS;QACX,CAAC;QAED,IAAI,CAAC,OAAO;YAAE,SAAS;QAEvB,MAAM,KAAK,GAAG,IAAI,CAAC,QAAQ,CAAC,GAAG,CAAC,QAAQ,EAAE,IAAI,CAAC,OAAO,CAAC,GAAG,CAAC,QAAQ,CAAC,CAAC,CAAC;QACtE,MAAM,GAAG,GAAG,IAAI,CAAC,OAAO,CAAC,GAAG,CAAC,QAAQ,CAAC,CAAC,WAAW,EAAE,CAAC;QACrD,MAAM,YAAY,GAAG,IAAI,CAAC,QAAQ,CAAC,OAAO,EAAE,GAAG,CAAC,QAAQ,CAAC,CAAC;QAC1D,MAAM,IAAI,GAAG,KAAK,CAAC,OAAO,CAAC,mBAAmB,EAAE,GAAG,CAAC,CAAC,WAAW,EAAE,CAAC;QACnE,MAAM,OAAO,GAAG,IAAI,CAAC,IAAI,CAAC,SAAS,EAAE,GAAG,IAAI,KAAK,CAAC,CAAC;QAEnD,MAAM,UAAU,GAAG,CAAC,CAAS,EAAE,EAAE,CAC/B,qBAAqB,CAAC,IAAI,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,IAAI,CAAC,CAAC,OAAO,CAAC,KAAK,EAAE,MAAM,CAAC,CAAC,OAAO,CAAC,IAAI,EAAE,KAAK,CAAC,GAAG,CAAC,CAAC,CAAC,CAAC,CAAC;QAC3F,MAAM,WAAW,GAAG;YAClB,KAAK;YACL,gBAAgB;YAChB,UAAU,UAAU,CAAC,KAAK,CAAC,EAAE;YAC7B,gBAAgB,UAAU,CAAC,kBAAkB,IAAI,CAAC,QAAQ,CAAC,GAAG,CAAC,QAAQ,CAAC,EAAE,CAAC,EAAE;YAC7E,aAAa,UAAU,CAAC,YAAY,CAAC,EAAE;YACvC,sBAAsB,UAAU,CAAC,GAAG,CAAC,KAAK,CAAC,CAAC,CAAC,CAAC,GAAG;YACjD,cAAc,UAAU,CAAC,GAAG,CAAC,SAAS,CAAC,EAAE;YACzC,gBAAgB,UAAU,CAAC,GAAG,CAAC,WAAW,CAAC,EAAE;YAC7C,eAAe,GAAG,CAAC,UAAU,EAAE;YAC/B,KAAK;YACL,EAAE;SACH,CAAC,IAAI,CAAC,IAAI,CAAC,CAAC;QAEb,EAAE,CAAC,aAAa,CAAC,OAAO,EAAE,WAAW,GAAG,OAAO,EAAE,OAAO,CAAC,CAAC;QAC1D,YAAY,CAAC,IAAI,CACf,MAAM,KAAK,KAAK,IAAI,UAAU,IAAI,CAAC,QAAQ,CAAC,GAAG,CAAC,QAAQ,CAAC,KAAK,GAAG,CAAC,UAAU,UAAU,CACvF,CAAC;QACF,QAAQ,EAAE,CAAC;IACb,CAAC;IAED,iBAAiB;IACjB,MAAM,YAAY,GAAG;QACnB,oBAAoB;QACpB,EAAE;QACF,6BAA6B,IAAI,IAAI,EAAE,CAAC,WAAW,EAAE,CAAC,KAAK,CAAC,CAAC,EAAE,EAAE,CAAC,EAAE;QACpE,EAAE;QACF,GAAG,YAAY;QACf,EAAE;KACH,CAAC,IAAI,CAAC,IAAI,CAAC,CAAC;IACb,EAAE,CAAC,aAAa,CAAC,IAAI,CAAC,IAAI,CAAC,SAAS,EAAE,UAAU,CAAC,EAAE,YAAY,EAAE,OAAO,CAAC,CAAC;IAE1E,OAAO,EAAE,QAAQ,EAAE,SAAS,EAAE,CAAC;AACjC,CAAC;AAED,sEAAsE;AAEtE,MAAM,CAAC,KAAK,UAAU,aAAa,CACjC,SAAiB,EACjB,KAAK,GAAG,QAAQ,EAChB,OAAO,GAAG,cAAc,EAAE;IAE1B,MAAM,QAAQ,GAAG,IAAI,CAAC,OAAO,CAAC,SAAS,CAAC,CAAC;IACzC,MAAM,KAAK,GAAG,EAAE;SACb,WAAW,CAAC,QAAQ,CAAC;SACrB,MAAM,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC,QAAQ,CAAC,KAAK,CAAC,IAAI,CAAC,KAAK,UAAU,IAAI,CAAC,KAAK,QAAQ,CAAC;SACtE,GAAG,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,IAAI,CAAC,IAAI,CAAC,QAAQ,EAAE,CAAC,CAAC,CAAC,CAAC;IAEtC,MAAM,MAAM,GAAsB;QAChC,cAAc,EAAE,CAAC;QACjB,YAAY,EAAE,CAAC;QACf,WAAW,EAAE,CAAC;QACd,MAAM,EAAE,EAAE;QACV,OAAO,EAAE,EAAE;KACZ,CAAC;IAEF,KAAK,MAAM,IAAI,IAAI,KAAK,EAAE,CAAC;QACzB,MAAM,CAAC,GAAG,MAAM,cAAc,CAAC,IAAI,EAAE,KAAK,EAAE,OAAO,CAAC,CAAC;QACrD,MAAM,CAAC,OAAO,CAAC,IAAI,CAAC,CAAC,CAAC,CAAC;QACvB,IAAI,CAAC,CAAC,OAAO,EAAE,CAAC;YACd,MAAM,CAAC,YAAY,EAAE,CAAC;QACxB,CAAC;aAAM,CAAC;YACN,MAAM,CAAC,cAAc,EAAE,CAAC;YACxB,MAAM,CAAC,WAAW,IAAI,CAAC,CAAC,aAAa,CAAC;QACxC,CAAC;QACD,IAAI,CAAC,CAAC,KAAK,IAAI,CAAC,CAAC,CAAC,OAAO;YAAE,MAAM,CAAC,MAAM,CAAC,IAAI,CAAC,GAAG,CAAC,CAAC,QAAQ,KAAK,CAAC,CAAC,KAAK,EAAE,CAAC,CAAC;IAC7E,CAAC;IAED,OAAO,MAAM,CAAC;AAChB,CAAC"}

package/dist/src/knowledge/profile-store.d.ts

@@ -0,0 +1,134 @@
+/**
+ * Per-profile capture stores (the browser track's profile feature).
+ *
+ * mono-agent has profiles — work, personal, a client — and a page captured
+ * from the browser can name the one it belongs to (`meta.profile` on the
+ * capture envelope). The promise that makes that worth having is negative:
+ * a page saved into `work` must NOT come back from a search of `personal`.
+ *
+ * WHICH MECHANISM, AND WHY. A store here is three things — a vector db
+ * directory (`dbPath`), a namespace, and the root the `doc-metadata.jsonl`
+ * lives under — and `scope` already decides all three (see
+ * document-pipeline: `namespace()`, `effectiveRoot()`, `storeDbPath()`).
+ * `global` is the existing proof: one scope value routes to a directory
+ * outside any project. A profile is the same shape, so it is expressed the
+ * same way — a scope of `profile:<id>` — rather than as a second, parallel
+ * "which brain" parameter threaded through ingest, search, cite and
+ * related. Nothing here is a new concept; it is one more value of an
+ * existing one.
+ *
+ * WHERE THE DIRECTORY GOES. Inside the global brain
+ * (`~/.monomind/global-brain/profiles/<id>`), not beside it. That is not
+ * cosmetic: `memory-bridge.getDbPath` refuses any dbPath outside the
+ * project, the per-project data dir, or the global brain, and — this is the
+ * dangerous part — a refused path does not throw, it silently falls back to
+ * the DEFAULT store. A profile brain anywhere else would quietly put every
+ * profile's captures in one database, which is the exact failure this
+ * feature exists to prevent. Being inside the global brain also inherits
+ * its protection from `cleanup --data`, which prunes `~/.monomind/projects`
+ * and deliberately never touches its sibling.
+ *
+ * UNTRUSTED INPUT. The profile id arrives from a browser extension and
+ * becomes a directory name. `isValidProfileId` is the guard, kept
+ * deliberately identical to Go's `profiledir.ValidProfileID` — the two
+ * sides of this contract must reject exactly the same strings.
+ *
+ * @module v1/cli/knowledge/profile-store
+ */
+/** Scopes of the form `profile:<id>` address one profile's own store. */
+export declare const PROFILE_SCOPE_PREFIX = "profile:";
+/**
+ * isValidProfileId mirrors `profiledir.ValidProfileID` (Go): non-empty, no
+ * path separator, no parent-directory component. Every other function here
+ * returns undefined for an id this rejects, so a hostile id can never
+ * become a path — not even a path that is then refused.
+ */
+export declare function isValidProfileId(id: unknown): id is string;
+/** The scope value addressing one profile's store, or undefined. */
+export declare function profileScope(id: unknown): string | undefined;
+/** The profile id inside a `profile:<id>` scope, or undefined for any other. */
+export declare function parseProfileScope(scope: unknown): string | undefined;
+/** True for a scope that names a profile. */
+export declare const isProfileScope: (scope: unknown) => boolean;
+/**
+ * profileBrainDir is a profile's store directory — the value used BOTH as
+ * the dbPath and as the metadata root, exactly as the global brain uses one
+ * directory for both.
+ *
+ * It creates the directory, which a resolver would not normally do. The
+ * reason is specific: `memory-bridge.getDbPath` compares
+ * `fs.realpathSync(dbPath)` against `fs.realpathSync(globalBrainDir)`, and
+ * a path that does not exist yet cannot be realpath'd. On a machine where
+ * the home directory is reached through a symlink, the two sides would then
+ * resolve differently, the guard would reject the path, and — because that
+ * guard falls back instead of throwing — the profile's captures would land
+ * in the default store together with every other profile's. Creating the
+ * directory first makes both sides resolvable and the comparison honest.
+ */
+export declare function profileBrainDir(id: unknown): string | undefined;
+/**
+ * profileStoreDir is profileBrainDir for a `profile:<id>` SCOPE, and
+ * undefined for every other scope — the single function
+ * `document-store`'s `effectiveRoot`/`storeDbPath` consult, so that the
+ * routing rule lives in one place and the pipeline keeps a two-line hook.
+ *
+ * DO NOT MOVE THIS DIRECTORY OUT OF THE GLOBAL BRAIN. The nesting looks
+ * arbitrary and is not: `memory-bridge.getDbPath` accepts a dbPath only
+ * inside the project, the per-project data dir, or the global brain, and a
+ * path it does not accept is NOT rejected — it SILENTLY RETURNS THE DEFAULT
+ * STORE. So a profile brain sited anywhere else (beside the global brain,
+ * under `~/.monoagent`, under the profile's own root) does not fail: every
+ * profile's captures quietly land in one database, `work` turns up in a
+ * `personal` search, and nothing anywhere reports a problem. The feature
+ * fails by succeeding. `src/__tests__/profile-store-dbpath.test.ts` holds
+ * both halves of that — this path surviving, and an outside path silently
+ * becoming the default.
+ */
+export declare function profileStoreDir(scope: unknown): string | undefined;
+/**
+ * profileInboxDir is where that profile's captures land, as the Go side
+ * writes them: `~/.monoagent/profiles/<id>/.monomind/inbox`.
+ *
+ * Keyed by profile id alone. mono-agent lets a profile's root be moved
+ * (`profiles.root_dir`), and the capture inbox deliberately does NOT follow
+ * it on either side — see `ProfileInbox` in `internal/capture/profile.go`
+ * for why. This path and that one are the same rule written twice, which is
+ * the cost of the two repos agreeing about a directory at all.
+ */
+export declare function profileInboxDir(id: unknown): string | undefined;
+/**
+ * profileInboxes lists the profile inboxes that exist on this machine, for
+ * a sweep that wants to ingest every profile's captures rather than only
+ * the shared inbox. Missing directories are simply absent; nothing here
+ * throws.
+ */
+export declare function profileInboxes(): Array<{
+    id: string;
+    inbox: string;
+    scope: string;
+}>;
+/**
+ * captureScope is the routing rule, applied once per ingest: a capture
+ * whose envelope names a profile belongs to that profile's store, whatever
+ * scope the caller reached for.
+ *
+ * The envelope wins on purpose. The profile was chosen by a person at save
+ * time, in the browser, and every automatic caller (`doc ingest` on a path
+ * outside any project, the dashboard's watcher, an inbox sweep) picks its
+ * scope from where the FILE is rather than from what it holds. A caller
+ * that genuinely means a particular store says so by passing a
+ * `profile:<id>` scope, which is returned untouched.
+ */
+export declare function captureScope(scope: string, filePath: string): string;
+/**
+ * envelopeProfile reads the `profile` out of the capture envelope holding
+ * `filePath`, if there is one. Defensive on the same terms as the rest of
+ * capture-envelope: `meta.json` is written by a browser extension on a
+ * machine we do not control, so every failure — absent, oversized,
+ * truncated, wrong-typed, or an id that could escape a directory — is
+ * "this capture names no profile", never a thrown error.
+ */
+export declare function envelopeProfile(filePath: string): string | undefined;
+/** A human-readable name for a scope, for a listing or a search result. */
+export declare function describeScope(scope: string): string;
+//# sourceMappingURL=profile-store.d.ts.map

package/dist/src/knowledge/profile-store.d.ts.map

@@ -0,0 +1 @@
+{"version":3,"file":"profile-store.d.ts","sourceRoot":"","sources":["../../../src/knowledge/profile-store.ts"],"names":[],"mappings":"AAAA;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;GAoCG;AAMH,yEAAyE;AACzE,eAAO,MAAM,oBAAoB,aAAa,CAAC;AAiB/C;;;;;GAKG;AACH,wBAAgB,gBAAgB,CAAC,EAAE,EAAE,OAAO,GAAG,EAAE,IAAI,MAAM,CAO1D;AAED,oEAAoE;AACpE,wBAAgB,YAAY,CAAC,EAAE,EAAE,OAAO,GAAG,MAAM,GAAG,SAAS,CAE5D;AAED,gFAAgF;AAChF,wBAAgB,iBAAiB,CAAC,KAAK,EAAE,OAAO,GAAG,MAAM,GAAG,SAAS,CAIpE;AAED,6CAA6C;AAC7C,eAAO,MAAM,cAAc,UAAW,OAAO,KAAG,OAAiD,CAAC;AAKlG;;;;;;;;;;;;;;GAcG;AACH,wBAAgB,eAAe,CAAC,EAAE,EAAE,OAAO,GAAG,MAAM,GAAG,SAAS,CAa/D;AAED;;;;;;;;;;;;;;;;;GAiBG;AACH,wBAAgB,eAAe,CAAC,KAAK,EAAE,OAAO,GAAG,MAAM,GAAG,SAAS,CAGlE;AAED;;;;;;;;;GASG;AACH,wBAAgB,eAAe,CAAC,EAAE,EAAE,OAAO,GAAG,MAAM,GAAG,SAAS,CAG/D;AAED;;;;;GAKG;AACH,wBAAgB,cAAc,IAAI,KAAK,CAAC;IAAE,EAAE,EAAE,MAAM,CAAC;IAAC,KAAK,EAAE,MAAM,CAAC;IAAC,KAAK,EAAE,MAAM,CAAA;CAAE,CAAC,CAqBpF;AAED;;;;;;;;;;;GAWG;AACH,wBAAgB,YAAY,CAAC,KAAK,EAAE,MAAM,EAAE,QAAQ,EAAE,MAAM,GAAG,MAAM,CAIpE;AAKD;;;;;;;GAOG;AACH,wBAAgB,eAAe,CAAC,QAAQ,EAAE,MAAM,GAAG,MAAM,GAAG,SAAS,CAYpE;AAED,2EAA2E;AAC3E,wBAAgB,aAAa,CAAC,KAAK,EAAE,MAAM,GAAG,MAAM,CAGnD"}