@noodleseed/one 0.125.0 → 0.127.0

package/node_modules/@noodle-borg/knowledge-crawl/dist/firecrawl.js.map

@@ -0,0 +1 @@
+{"version":3,"file":"firecrawl.js","sourceRoot":"","sources":["../src/firecrawl.ts"],"names":[],"mappings":"AAAA;;;;;;;;;GASG;AACH,OAAO,EAAE,cAAc,EAAE,MAAM,wBAAwB,CAAC;AACxD,OAAO,EAAE,UAAU,EAAE,MAAM,aAAa,CAAC;AACzC,OAAO,EAAE,WAAW,EAAE,MAAM,aAAa,CAAC;AAG1C,MAAM,GAAG,GAAG,oCAAoC,CAAC;AAcjD,MAAM,OAAO,oBAAoB;IACtB,QAAQ,CAEf;IAEF,YAAY,OAAoC;QAC9C,IAAI,CAAC,QAAQ,GAAG;YACd,MAAM,EAAE,OAAO,CAAC,MAAM;YACtB,SAAS,EAAE,OAAO,CAAC,SAAS,IAAI,KAAK;YACrC,cAAc,EAAE,OAAO,CAAC,cAAc,IAAI,IAAI;YAC9C,QAAQ,EAAE,OAAO,CAAC,QAAQ,IAAI,GAAG;SAClC,CAAC;IACJ,CAAC;IAED,KAAK,CAAC,SAAS,CAAC,OAAyB;QACvC,MAAM,MAAM,GAAG,WAAW,CAAC,OAAO,CAAC,MAAM,CAAC,CAAC;QAC3C,MAAM,OAAO,GAAG,CAAC,MAAM,IAAI,CAAC,KAAK,CAAC,OAAO,EAAE,GAAG,EAAE;YAC9C,MAAM,EAAE,MAAM;YACd,IAAI,EAAE,IAAI,CAAC,SAAS,CAAC;gBACnB,GAAG,EAAE,MAAM,CAAC,MAAM;gBAClB,KAAK,EAAE,OAAO,CAAC,QAAQ;gBACvB,aAAa,EAAE,EAAE,OAAO,EAAE,CAAC,EAAE,IAAI,EAAE,UAAU,EAAE,CAAC,EAAE;aACnD,CAAC;SACH,CAAC,CAAoB,CAAC;QACvB,IAAI,OAAO,OAAO,CAAC,EAAE,KAAK,QAAQ,IAAI,OAAO,CAAC,EAAE,KAAK,EAAE,EAAE,CAAC;YACxD,MAAM,IAAI,cAAc,CAAC,UAAU,EAAE,0CAA0C,CAAC,CAAC;QACnF,CAAC;QAED,MAAM,SAAS,GAAwB,EAAE,CAAC;QAC1C,IAAI,OAAO,GAAG,GAAG,GAAG,IAAI,OAAO,CAAC,EAAE,EAAE,CAAC;QACrC,KAAK,IAAI,IAAI,GAAG,CAAC,EAAE,IAAI,GAAG,IAAI,CAAC,QAAQ,CAAC,QAAQ,EAAE,IAAI,EAAE,EAAE,CAAC;YACzD,MAAM,MAAM,GAAG,CAAC,MAAM,IAAI,CAAC,KAAK,CAAC,QAAQ,EAAE,OAAO,CAAC,CAIlD,CAAC;YACF,IAAI,MAAM,CAAC,MAAM,KAAK,QAAQ,EAAE,CAAC;gBAC/B,MAAM,IAAI,cAAc,CAAC,UAAU,EAAE,wBAAwB,CAAC,CAAC;YACjE,CAAC;YACD,IAAI,MAAM,CAAC,MAAM,KAAK,WAAW,EAAE,CAAC;gBAClC,SAAS,CAAC,IAAI,CAAC,GAAG,CAAC,MAAM,CAAC,IAAI,IAAI,EAAE,CAAC,CAAC,CAAC;gBACvC,IAAI,OAAO,MAAM,CAAC,IAAI,KAAK,QAAQ,IAAI,MAAM,CAAC,IAAI,KAAK,EAAE,EAAE,CAAC;oBAC1D,OAAO,GAAG,MAAM,CAAC,IAAI,CAAC;oBACtB,SAAS;gBACX,CAAC;gBACD,OAAO,UAAU,CACf,SAAS,CAAC,GAAG,CAAC,CAAC,QAAQ,EAAE,EAAE,CAAC,CAAC;oBAC3B,GAAG,EAAE,QAAQ,CAAC,QAAQ,EAAE,SAAS,IAAI,EAAE;oBACvC,KAAK,EAAE,cAAc,CAAC,QAAQ,CAAC,QAAQ,EAAE,KAAK,CAAC;oBAC/C,IAAI,EAAE,QAAQ,CAAC,QAAQ,IAAI,EAAE;iBAC9B,CAAC,CAAC,EACH,OAAO,EACP,MAAM,CACP,CAAC;YACJ,CAAC;YACD,IAAI,IAAI,CAAC,QAAQ,CAAC,cAAc,GAAG,CAAC,EAAE,CAAC;gBACrC,MAAM,IAAI,OAAO,CAAC,CAAC,OAAO,EAAE,EAAE,CAAC,UAAU,CAAC,OAAO,EAAE,IAAI,CAAC,QAAQ,CAAC,cAAc,CAAC,CAAC,CAAC;YACpF,CAAC;QACH,CAAC;QACD,MAAM,IAAI,cAAc,CAAC,UAAU,EAAE,yDAAyD,CAAC,CAAC;IAClG,CAAC;IAED,KAAK,CAAC,KAAK,CAAC,SAAiB,EAAE,GAAW,EAAE,OAAoB,EAAE;QAChE,IAAI,QAAkB,CAAC;QACvB,IAAI,CAAC;YACH,QAAQ,GAAG,MAAM,IAAI,CAAC,QAAQ,CAAC,SAAS,CAAC,GAAG,EAAE;gBAC5C,GAAG,IAAI;gBACP,OAAO,EAAE;oBACP,aAAa,EAAE,UAAU,IAAI,CAAC,QAAQ,CAAC,MAAM,EAAE;oBAC/C,cAAc,EAAE,kBAAkB;iBACnC;aACF,CAAC,CAAC;QACL,CAAC;QAAC,MAAM,CAAC;YACP,MAAM,IAAI,cAAc,CAAC,UAAU,EAAE,aAAa,SAAS,iBAAiB,CAAC,CAAC;QAChF,CAAC;QACD,IAAI,CAAC,QAAQ,CAAC,EAAE,EAAE,CAAC;YACjB,MAAM,IAAI,cAAc,CACtB,UAAU,EACV,aAAa,SAAS,kBAAkB,QAAQ,CAAC,MAAM,EAAE,CAC1D,CAAC;QACJ,CAAC;QACD,OAAO,QAAQ,CAAC,IAAI,EAAE,CAAC,KAAK,CAAC,GAAG,EAAE,CAAC,CAAC,EAAE,CAAC,CAAC,CAAC;IAC3C,CAAC;CACF;AAED,SAAS,cAAc,CAAC,KAA6C;IACnE,IAAI,OAAO,KAAK,KAAK,QAAQ;QAAE,OAAO,KAAK,CAAC;IAC5C,IAAI,KAAK,CAAC,OAAO,CAAC,KAAK,CAAC;QAAE,OAAO,KAAK,CAAC,IAAI,CAAC,CAAC,KAAK,EAAE,EAAE,CAAC,OAAO,KAAK,KAAK,QAAQ,CAAC,IAAI,EAAE,CAAC;IACxF,OAAO,EAAE,CAAC;AACZ,CAAC"}

package/node_modules/@noodle-borg/knowledge-crawl/dist/first-party.d.ts

@@ -0,0 +1,11 @@
+import type { CrawlPage, SiteCrawlRequest, SiteFetcher } from './ports.js';
+export interface FirstPartySiteFetcherOptions {
+    /** Injected for tests; defaults to the SSRF-guarded fetch. */
+    readonly fetchImpl?: typeof fetch;
+}
+export declare class FirstPartySiteFetcher implements SiteFetcher {
+    #private;
+    constructor(options?: FirstPartySiteFetcherOptions);
+    fetchSite(request: SiteCrawlRequest): Promise<readonly CrawlPage[]>;
+}
+//# sourceMappingURL=first-party.d.ts.map

package/node_modules/@noodle-borg/knowledge-crawl/dist/first-party.d.ts.map

@@ -0,0 +1 @@
+{"version":3,"file":"first-party.d.ts","sourceRoot":"","sources":["../src/first-party.ts"],"names":[],"mappings":"AAcA,OAAO,KAAK,EAAE,SAAS,EAAE,gBAAgB,EAAE,WAAW,EAAE,MAAM,YAAY,CAAC;AAO3E,MAAM,WAAW,4BAA4B;IAC3C,8DAA8D;IAC9D,QAAQ,CAAC,SAAS,CAAC,EAAE,OAAO,KAAK,CAAC;CACnC;AAED,qBAAa,qBAAsB,YAAW,WAAW;;gBAG3C,OAAO,GAAE,4BAAiC;IAMhD,SAAS,CAAC,OAAO,EAAE,gBAAgB,GAAG,OAAO,CAAC,SAAS,SAAS,EAAE,CAAC;CAkF1E"}

package/node_modules/@noodle-borg/knowledge-crawl/dist/first-party.js

@@ -0,0 +1,108 @@
+/**
+ * The managed first-party site fetcher (ADR 0202 amendment 2026-08-18): sitemap-first discovery,
+ * an include-contained same-origin link walk, robots.txt, and bounded, deterministic output.
+ *
+ * Egress is SSRF-guarded in layers: the default fetch path is `guardedFetch` from
+ * `@noodle-borg/connector-http` (socket-level DNS pinning to public unicast addresses, ADR 0026),
+ * and this fetcher additionally enforces HTTPS-only, rejects literal-IP origins (no DNS step means
+ * no pinning), keeps every fetched URL on the declared origin, and never follows redirects
+ * (a moved page is indexed at its new URL when discovery finds it there).
+ */
+import { guardedFetch } from '@noodle-borg/connector-http';
+import { pathMatches } from '@noodle-borg/knowledge';
+import { extractPageContent } from './html-text.js';
+import { parseOrigin, safeUrl } from './origin.js';
+import { parseRobots } from './robots.js';
+import { parseSitemapLocations } from './sitemap.js';
+const ALLOW_ALL = { isAllowed: () => true, sitemaps: [] };
+const LINK = /<a\s[^>]*?href\s*=\s*["']([^"'#]+)["']/gi;
+export class FirstPartySiteFetcher {
+    #fetch;
+    constructor(options = {}) {
+        this.#fetch =
+            options.fetchImpl ??
+                ((url, init) => guardedFetch(new URL(String(url)), init));
+    }
+    async fetchSite(request) {
+        const origin = parseOrigin(request.origin);
+        const robots = await this.#robots(origin);
+        const indexable = (url) => url.origin === origin.origin &&
+            robots.isAllowed(url.pathname) &&
+            request.include.some((glob) => pathMatches(url.pathname, glob));
+        const queue = new Set();
+        for (const location of await this.#sitemapLocations(origin, robots)) {
+            const url = safeUrl(location);
+            if (url !== undefined && indexable(url))
+                queue.add(url.origin + url.pathname);
+        }
+        const pages = [];
+        const visited = new Set();
+        // The root is fetched for discovery even when it is not itself indexable.
+        const discovery = [`${origin.origin}/`];
+        for (;;) {
+            const next = [...queue].filter((url) => !visited.has(url)).sort()[0] ??
+                discovery.find((url) => !visited.has(url));
+            if (next === undefined || pages.length >= request.maxPages)
+                break;
+            visited.add(next);
+            const fetched = await this.#page(next, request.maxPageBytes);
+            if (fetched === undefined)
+                continue;
+            for (const href of fetched.html.matchAll(LINK)) {
+                const linked = safeUrl(href[1] ?? '', next);
+                if (linked !== undefined && indexable(linked))
+                    queue.add(linked.origin + linked.pathname);
+            }
+            if (indexable(new URL(next)) && fetched.text !== '') {
+                pages.push({
+                    url: next,
+                    title: fetched.title === '' ? next : fetched.title,
+                    text: fetched.text,
+                });
+            }
+        }
+        return pages.sort((a, b) => a.url.localeCompare(b.url));
+    }
+    async #robots(origin) {
+        try {
+            const response = await this.#fetch(`${origin.origin}/robots.txt`);
+            if (response.status !== 200)
+                return ALLOW_ALL;
+            return parseRobots(await response.text());
+        }
+        catch {
+            return ALLOW_ALL;
+        }
+    }
+    async #sitemapLocations(origin, robots) {
+        const sitemapUrls = (robots.sitemaps.length > 0 ? robots.sitemaps : [`${origin.origin}/sitemap.xml`]).filter((url) => safeUrl(url)?.origin === origin.origin);
+        const locations = [];
+        for (const sitemapUrl of sitemapUrls) {
+            try {
+                const response = await this.#fetch(sitemapUrl);
+                if (response.status !== 200)
+                    continue;
+                locations.push(...parseSitemapLocations(await response.text()));
+            }
+            catch {
+                // A missing or failing sitemap is not an error; the link walk still discovers pages.
+            }
+        }
+        return locations;
+    }
+    async #page(url, maxPageBytes) {
+        try {
+            const response = await this.#fetch(url, { redirect: 'manual' });
+            if (response.status !== 200)
+                return undefined;
+            const bytes = await response.arrayBuffer();
+            if (bytes.byteLength > maxPageBytes)
+                return undefined;
+            return extractPageContent(response.headers.get('content-type') ?? '', bytes);
+        }
+        catch {
+            return undefined;
+        }
+    }
+}
+//# sourceMappingURL=first-party.js.map

package/node_modules/@noodle-borg/knowledge-crawl/dist/first-party.js.map

@@ -0,0 +1 @@
+{"version":3,"file":"first-party.js","sourceRoot":"","sources":["../src/first-party.ts"],"names":[],"mappings":"AAAA;;;;;;;;;GASG;AACH,OAAO,EAAE,YAAY,EAAE,MAAM,6BAA6B,CAAC;AAC3D,OAAO,EAAE,WAAW,EAAE,MAAM,wBAAwB,CAAC;AACrD,OAAO,EAAE,kBAAkB,EAAE,MAAM,gBAAgB,CAAC;AACpD,OAAO,EAAE,WAAW,EAAE,OAAO,EAAE,MAAM,aAAa,CAAC;AAEnD,OAAO,EAAE,WAAW,EAAe,MAAM,aAAa,CAAC;AACvD,OAAO,EAAE,qBAAqB,EAAE,MAAM,cAAc,CAAC;AAErD,MAAM,SAAS,GAAW,EAAE,SAAS,EAAE,GAAG,EAAE,CAAC,IAAI,EAAE,QAAQ,EAAE,EAAE,EAAE,CAAC;AAClE,MAAM,IAAI,GAAG,0CAA0C,CAAC;AAOxD,MAAM,OAAO,qBAAqB;IACvB,MAAM,CAAe;IAE9B,YAAY,UAAwC,EAAE;QACpD,IAAI,CAAC,MAAM;YACT,OAAO,CAAC,SAAS;gBAChB,CAAC,CAAC,GAAG,EAAE,IAAI,EAAE,EAAE,CAAC,YAAY,CAAC,IAAI,GAAG,CAAC,MAAM,CAAC,GAAG,CAAC,CAAC,EAAE,IAAmB,CAAC,CAAkB,CAAC;IAC/F,CAAC;IAED,KAAK,CAAC,SAAS,CAAC,OAAyB;QACvC,MAAM,MAAM,GAAG,WAAW,CAAC,OAAO,CAAC,MAAM,CAAC,CAAC;QAC3C,MAAM,MAAM,GAAG,MAAM,IAAI,CAAC,OAAO,CAAC,MAAM,CAAC,CAAC;QAC1C,MAAM,SAAS,GAAG,CAAC,GAAQ,EAAW,EAAE,CACtC,GAAG,CAAC,MAAM,KAAK,MAAM,CAAC,MAAM;YAC5B,MAAM,CAAC,SAAS,CAAC,GAAG,CAAC,QAAQ,CAAC;YAC9B,OAAO,CAAC,OAAO,CAAC,IAAI,CAAC,CAAC,IAAI,EAAE,EAAE,CAAC,WAAW,CAAC,GAAG,CAAC,QAAQ,EAAE,IAAI,CAAC,CAAC,CAAC;QAElE,MAAM,KAAK,GAAG,IAAI,GAAG,EAAU,CAAC;QAChC,KAAK,MAAM,QAAQ,IAAI,MAAM,IAAI,CAAC,iBAAiB,CAAC,MAAM,EAAE,MAAM,CAAC,EAAE,CAAC;YACpE,MAAM,GAAG,GAAG,OAAO,CAAC,QAAQ,CAAC,CAAC;YAC9B,IAAI,GAAG,KAAK,SAAS,IAAI,SAAS,CAAC,GAAG,CAAC;gBAAE,KAAK,CAAC,GAAG,CAAC,GAAG,CAAC,MAAM,GAAG,GAAG,CAAC,QAAQ,CAAC,CAAC;QAChF,CAAC;QAED,MAAM,KAAK,GAAgB,EAAE,CAAC;QAC9B,MAAM,OAAO,GAAG,IAAI,GAAG,EAAU,CAAC;QAClC,0EAA0E;QAC1E,MAAM,SAAS,GAAG,CAAC,GAAG,MAAM,CAAC,MAAM,GAAG,CAAC,CAAC;QACxC,SAAS,CAAC;YACR,MAAM,IAAI,GACR,CAAC,GAAG,KAAK,CAAC,CAAC,MAAM,CAAC,CAAC,GAAG,EAAE,EAAE,CAAC,CAAC,OAAO,CAAC,GAAG,CAAC,GAAG,CAAC,CAAC,CAAC,IAAI,EAAE,CAAC,CAAC,CAAC;gBACvD,SAAS,CAAC,IAAI,CAAC,CAAC,GAAG,EAAE,EAAE,CAAC,CAAC,OAAO,CAAC,GAAG,CAAC,GAAG,CAAC,CAAC,CAAC;YAC7C,IAAI,IAAI,KAAK,SAAS,IAAI,KAAK,CAAC,MAAM,IAAI,OAAO,CAAC,QAAQ;gBAAE,MAAM;YAClE,OAAO,CAAC,GAAG,CAAC,IAAI,CAAC,CAAC;YAClB,MAAM,OAAO,GAAG,MAAM,IAAI,CAAC,KAAK,CAAC,IAAI,EAAE,OAAO,CAAC,YAAY,CAAC,CAAC;YAC7D,IAAI,OAAO,KAAK,SAAS;gBAAE,SAAS;YACpC,KAAK,MAAM,IAAI,IAAI,OAAO,CAAC,IAAI,CAAC,QAAQ,CAAC,IAAI,CAAC,EAAE,CAAC;gBAC/C,MAAM,MAAM,GAAG,OAAO,CAAC,IAAI,CAAC,CAAC,CAAC,IAAI,EAAE,EAAE,IAAI,CAAC,CAAC;gBAC5C,IAAI,MAAM,KAAK,SAAS,IAAI,SAAS,CAAC,MAAM,CAAC;oBAAE,KAAK,CAAC,GAAG,CAAC,MAAM,CAAC,MAAM,GAAG,MAAM,CAAC,QAAQ,CAAC,CAAC;YAC5F,CAAC;YACD,IAAI,SAAS,CAAC,IAAI,GAAG,CAAC,IAAI,CAAC,CAAC,IAAI,OAAO,CAAC,IAAI,KAAK,EAAE,EAAE,CAAC;gBACpD,KAAK,CAAC,IAAI,CAAC;oBACT,GAAG,EAAE,IAAI;oBACT,KAAK,EAAE,OAAO,CAAC,KAAK,KAAK,EAAE,CAAC,CAAC,CAAC,IAAI,CAAC,CAAC,CAAC,OAAO,CAAC,KAAK;oBAClD,IAAI,EAAE,OAAO,CAAC,IAAI;iBACnB,CAAC,CAAC;YACL,CAAC;QACH,CAAC;QACD,OAAO,KAAK,CAAC,IAAI,CAAC,CAAC,CAAC,EAAE,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC,GAAG,CAAC,aAAa,CAAC,CAAC,CAAC,GAAG,CAAC,CAAC,CAAC;IAC1D,CAAC;IAED,KAAK,CAAC,OAAO,CAAC,MAAW;QACvB,IAAI,CAAC;YACH,MAAM,QAAQ,GAAG,MAAM,IAAI,CAAC,MAAM,CAAC,GAAG,MAAM,CAAC,MAAM,aAAa,CAAC,CAAC;YAClE,IAAI,QAAQ,CAAC,MAAM,KAAK,GAAG;gBAAE,OAAO,SAAS,CAAC;YAC9C,OAAO,WAAW,CAAC,MAAM,QAAQ,CAAC,IAAI,EAAE,CAAC,CAAC;QAC5C,CAAC;QAAC,MAAM,CAAC;YACP,OAAO,SAAS,CAAC;QACnB,CAAC;IACH,CAAC;IAED,KAAK,CAAC,iBAAiB,CAAC,MAAW,EAAE,MAAc;QACjD,MAAM,WAAW,GAAG,CAClB,MAAM,CAAC,QAAQ,CAAC,MAAM,GAAG,CAAC,CAAC,CAAC,CAAC,MAAM,CAAC,QAAQ,CAAC,CAAC,CAAC,CAAC,GAAG,MAAM,CAAC,MAAM,cAAc,CAAC,CAChF,CAAC,MAAM,CAAC,CAAC,GAAG,EAAE,EAAE,CAAC,OAAO,CAAC,GAAG,CAAC,EAAE,MAAM,KAAK,MAAM,CAAC,MAAM,CAAC,CAAC;QAC1D,MAAM,SAAS,GAAa,EAAE,CAAC;QAC/B,KAAK,MAAM,UAAU,IAAI,WAAW,EAAE,CAAC;YACrC,IAAI,CAAC;gBACH,MAAM,QAAQ,GAAG,MAAM,IAAI,CAAC,MAAM,CAAC,UAAU,CAAC,CAAC;gBAC/C,IAAI,QAAQ,CAAC,MAAM,KAAK,GAAG;oBAAE,SAAS;gBACtC,SAAS,CAAC,IAAI,CAAC,GAAG,qBAAqB,CAAC,MAAM,QAAQ,CAAC,IAAI,EAAE,CAAC,CAAC,CAAC;YAClE,CAAC;YAAC,MAAM,CAAC;gBACP,qFAAqF;YACvF,CAAC;QACH,CAAC;QACD,OAAO,SAAS,CAAC;IACnB,CAAC;IAED,KAAK,CAAC,KAAK,CACT,GAAW,EACX,YAAoB;QAEpB,IAAI,CAAC;YACH,MAAM,QAAQ,GAAG,MAAM,IAAI,CAAC,MAAM,CAAC,GAAG,EAAE,EAAE,QAAQ,EAAE,QAAQ,EAAiB,CAAC,CAAC;YAC/E,IAAI,QAAQ,CAAC,MAAM,KAAK,GAAG;gBAAE,OAAO,SAAS,CAAC;YAC9C,MAAM,KAAK,GAAG,MAAM,QAAQ,CAAC,WAAW,EAAE,CAAC;YAC3C,IAAI,KAAK,CAAC,UAAU,GAAG,YAAY;gBAAE,OAAO,SAAS,CAAC;YACtD,OAAO,kBAAkB,CAAC,QAAQ,CAAC,OAAO,CAAC,GAAG,CAAC,cAAc,CAAC,IAAI,EAAE,EAAE,KAAK,CAAC,CAAC;QAC/E,CAAC;QAAC,MAAM,CAAC;YACP,OAAO,SAAS,CAAC;QACnB,CAAC;IACH,CAAC;CACF"}

package/node_modules/@noodle-borg/knowledge-crawl/dist/html-text.d.ts

@@ -0,0 +1,19 @@
+/**
+ * Deliberately small HTML→text extraction for a lexical index: strip non-content blocks and
+ * markup, decode the common entities, collapse whitespace. Fetched pages are untrusted input —
+ * this reduces them to bounded plain text and nothing here evaluates or resolves anything.
+ */
+export declare function extractHtmlText(html: string): string;
+/**
+ * The content-extraction seam: (content type, bytes) → indexable `{ html, title, text }`, or
+ * `undefined` for media we cannot yet extract. Future formats — PDF is the named next candidate
+ * (ADR 0202 deferred triggers) — become one new case here; everything downstream (sealing,
+ * revisions, indexing, fusion, citations) is format-agnostic.
+ */
+export declare function extractPageContent(contentType: string, bytes: ArrayBuffer): {
+    html: string;
+    title: string;
+    text: string;
+} | undefined;
+export declare function extractHtmlTitle(html: string): string;
+//# sourceMappingURL=html-text.d.ts.map

package/node_modules/@noodle-borg/knowledge-crawl/dist/html-text.d.ts.map

@@ -0,0 +1 @@
+{"version":3,"file":"html-text.d.ts","sourceRoot":"","sources":["../src/html-text.ts"],"names":[],"mappings":"AAAA;;;;GAIG;AAgCH,wBAAgB,eAAe,CAAC,IAAI,EAAE,MAAM,GAAG,MAAM,CAKpD;AAED;;;;;GAKG;AACH,wBAAgB,kBAAkB,CAChC,WAAW,EAAE,MAAM,EACnB,KAAK,EAAE,WAAW,GACjB;IAAE,IAAI,EAAE,MAAM,CAAC;IAAC,KAAK,EAAE,MAAM,CAAC;IAAC,IAAI,EAAE,MAAM,CAAA;CAAE,GAAG,SAAS,CAI3D;AAED,wBAAgB,gBAAgB,CAAC,IAAI,EAAE,MAAM,GAAG,MAAM,CAYrD"}

package/node_modules/@noodle-borg/knowledge-crawl/dist/html-text.js

@@ -0,0 +1,62 @@
+/**
+ * Deliberately small HTML→text extraction for a lexical index: strip non-content blocks and
+ * markup, decode the common entities, collapse whitespace. Fetched pages are untrusted input —
+ * this reduces them to bounded plain text and nothing here evaluates or resolves anything.
+ */
+const NON_CONTENT_BLOCKS = /<(script|style|noscript|template|svg)\b[\s\S]*?<\/\1\s*>/gi;
+const BLOCK_BREAK = /<\/?(p|div|li|ul|ol|table|tr|td|th|h[1-6]|br|hr|section|article|header|footer|nav|blockquote|pre)\b[^>]*>/gi;
+const NAMED_ENTITIES = {
+    amp: '&',
+    lt: '<',
+    gt: '>',
+    quot: '"',
+    apos: "'",
+    nbsp: ' ',
+    mdash: '—',
+    ndash: '–',
+    hellip: '…',
+    rsquo: '’',
+    lsquo: '‘',
+    rdquo: '”',
+    ldquo: '“',
+    copy: '©',
+};
+function decodeEntities(text) {
+    return text
+        .replace(/&#x([0-9a-f]+);/gi, (_, hex) => String.fromCodePoint(Number.parseInt(hex, 16)))
+        .replace(/&#(\d+);/g, (_, dec) => String.fromCodePoint(Number.parseInt(dec, 10)))
+        .replace(/&([a-z]+);/gi, (match, name) => NAMED_ENTITIES[name.toLowerCase()] ?? match);
+}
+export function extractHtmlText(html) {
+    const withoutBlocks = html.replace(NON_CONTENT_BLOCKS, ' ');
+    const withBreaks = withoutBlocks.replace(BLOCK_BREAK, ' ');
+    const withoutTags = withBreaks.replace(/<[^>]+>/g, '');
+    return decodeEntities(withoutTags).replace(/\s+/g, ' ').trim();
+}
+/**
+ * The content-extraction seam: (content type, bytes) → indexable `{ html, title, text }`, or
+ * `undefined` for media we cannot yet extract. Future formats — PDF is the named next candidate
+ * (ADR 0202 deferred triggers) — become one new case here; everything downstream (sealing,
+ * revisions, indexing, fusion, citations) is format-agnostic.
+ */
+export function extractPageContent(contentType, bytes) {
+    if (!contentType.startsWith('text/html'))
+        return undefined;
+    const html = new TextDecoder().decode(bytes);
+    return { html, title: extractHtmlTitle(html), text: extractHtmlText(html) };
+}
+export function extractHtmlTitle(html) {
+    const title = /<title[^>]*>([\s\S]*?)<\/title\s*>/i.exec(html)?.[1];
+    if (title !== undefined && title.trim() !== '')
+        return decodeEntities(title).replace(/\s+/g, ' ').trim();
+    const heading = /<h1[^>]*>([\s\S]*?)<\/h1\s*>/i.exec(html)?.[1];
+    if (heading !== undefined) {
+        const text = decodeEntities(heading.replace(/<[^>]+>/g, ''))
+            .replace(/\s+/g, ' ')
+            .trim();
+        if (text !== '')
+            return text;
+    }
+    return '';
+}
+//# sourceMappingURL=html-text.js.map

package/node_modules/@noodle-borg/knowledge-crawl/dist/html-text.js.map

@@ -0,0 +1 @@
+{"version":3,"file":"html-text.js","sourceRoot":"","sources":["../src/html-text.ts"],"names":[],"mappings":"AAAA;;;;GAIG;AAEH,MAAM,kBAAkB,GAAG,4DAA4D,CAAC;AACxF,MAAM,WAAW,GACf,6GAA6G,CAAC;AAEhH,MAAM,cAAc,GAAqC;IACvD,GAAG,EAAE,GAAG;IACR,EAAE,EAAE,GAAG;IACP,EAAE,EAAE,GAAG;IACP,IAAI,EAAE,GAAG;IACT,IAAI,EAAE,GAAG;IACT,IAAI,EAAE,GAAG;IACT,KAAK,EAAE,GAAG;IACV,KAAK,EAAE,GAAG;IACV,MAAM,EAAE,GAAG;IACX,KAAK,EAAE,GAAG;IACV,KAAK,EAAE,GAAG;IACV,KAAK,EAAE,GAAG;IACV,KAAK,EAAE,GAAG;IACV,IAAI,EAAE,GAAG;CACV,CAAC;AAEF,SAAS,cAAc,CAAC,IAAY;IAClC,OAAO,IAAI;SACR,OAAO,CAAC,mBAAmB,EAAE,CAAC,CAAC,EAAE,GAAW,EAAE,EAAE,CAC/C,MAAM,CAAC,aAAa,CAAC,MAAM,CAAC,QAAQ,CAAC,GAAG,EAAE,EAAE,CAAC,CAAC,CAC/C;SACA,OAAO,CAAC,WAAW,EAAE,CAAC,CAAC,EAAE,GAAW,EAAE,EAAE,CAAC,MAAM,CAAC,aAAa,CAAC,MAAM,CAAC,QAAQ,CAAC,GAAG,EAAE,EAAE,CAAC,CAAC,CAAC;SACxF,OAAO,CAAC,cAAc,EAAE,CAAC,KAAK,EAAE,IAAY,EAAE,EAAE,CAAC,cAAc,CAAC,IAAI,CAAC,WAAW,EAAE,CAAC,IAAI,KAAK,CAAC,CAAC;AACnG,CAAC;AAED,MAAM,UAAU,eAAe,CAAC,IAAY;IAC1C,MAAM,aAAa,GAAG,IAAI,CAAC,OAAO,CAAC,kBAAkB,EAAE,GAAG,CAAC,CAAC;IAC5D,MAAM,UAAU,GAAG,aAAa,CAAC,OAAO,CAAC,WAAW,EAAE,GAAG,CAAC,CAAC;IAC3D,MAAM,WAAW,GAAG,UAAU,CAAC,OAAO,CAAC,UAAU,EAAE,EAAE,CAAC,CAAC;IACvD,OAAO,cAAc,CAAC,WAAW,CAAC,CAAC,OAAO,CAAC,MAAM,EAAE,GAAG,CAAC,CAAC,IAAI,EAAE,CAAC;AACjE,CAAC;AAED;;;;;GAKG;AACH,MAAM,UAAU,kBAAkB,CAChC,WAAmB,EACnB,KAAkB;IAElB,IAAI,CAAC,WAAW,CAAC,UAAU,CAAC,WAAW,CAAC;QAAE,OAAO,SAAS,CAAC;IAC3D,MAAM,IAAI,GAAG,IAAI,WAAW,EAAE,CAAC,MAAM,CAAC,KAAK,CAAC,CAAC;IAC7C,OAAO,EAAE,IAAI,EAAE,KAAK,EAAE,gBAAgB,CAAC,IAAI,CAAC,EAAE,IAAI,EAAE,eAAe,CAAC,IAAI,CAAC,EAAE,CAAC;AAC9E,CAAC;AAED,MAAM,UAAU,gBAAgB,CAAC,IAAY;IAC3C,MAAM,KAAK,GAAG,qCAAqC,CAAC,IAAI,CAAC,IAAI,CAAC,EAAE,CAAC,CAAC,CAAC,CAAC;IACpE,IAAI,KAAK,KAAK,SAAS,IAAI,KAAK,CAAC,IAAI,EAAE,KAAK,EAAE;QAC5C,OAAO,cAAc,CAAC,KAAK,CAAC,CAAC,OAAO,CAAC,MAAM,EAAE,GAAG,CAAC,CAAC,IAAI,EAAE,CAAC;IAC3D,MAAM,OAAO,GAAG,+BAA+B,CAAC,IAAI,CAAC,IAAI,CAAC,EAAE,CAAC,CAAC,CAAC,CAAC;IAChE,IAAI,OAAO,KAAK,SAAS,EAAE,CAAC;QAC1B,MAAM,IAAI,GAAG,cAAc,CAAC,OAAO,CAAC,OAAO,CAAC,UAAU,EAAE,EAAE,CAAC,CAAC;aACzD,OAAO,CAAC,MAAM,EAAE,GAAG,CAAC;aACpB,IAAI,EAAE,CAAC;QACV,IAAI,IAAI,KAAK,EAAE;YAAE,OAAO,IAAI,CAAC;IAC/B,CAAC;IACD,OAAO,EAAE,CAAC;AACZ,CAAC"}

package/node_modules/@noodle-borg/knowledge-crawl/dist/index.d.ts

@@ -0,0 +1,5 @@
+export { FirecrawlSiteFetcher, type FirecrawlSiteFetcherOptions } from './firecrawl.js';
+export { FirstPartySiteFetcher, type FirstPartySiteFetcherOptions } from './first-party.js';
+export type { CrawlPage, SiteCrawlRequest, SiteFetcher } from './ports.js';
+export { TavilySiteFetcher, type TavilySiteFetcherOptions } from './tavily.js';
+//# sourceMappingURL=index.d.ts.map

package/node_modules/@noodle-borg/knowledge-crawl/dist/index.d.ts.map

@@ -0,0 +1 @@
+{"version":3,"file":"index.d.ts","sourceRoot":"","sources":["../src/index.ts"],"names":[],"mappings":"AAAA,OAAO,EAAE,oBAAoB,EAAE,KAAK,2BAA2B,EAAE,MAAM,gBAAgB,CAAC;AACxF,OAAO,EAAE,qBAAqB,EAAE,KAAK,4BAA4B,EAAE,MAAM,kBAAkB,CAAC;AAC5F,YAAY,EAAE,SAAS,EAAE,gBAAgB,EAAE,WAAW,EAAE,MAAM,YAAY,CAAC;AAC3E,OAAO,EAAE,iBAAiB,EAAE,KAAK,wBAAwB,EAAE,MAAM,aAAa,CAAC"}

package/node_modules/@noodle-borg/knowledge-crawl/dist/index.js

@@ -0,0 +1,4 @@
+export { FirecrawlSiteFetcher } from './firecrawl.js';
+export { FirstPartySiteFetcher } from './first-party.js';
+export { TavilySiteFetcher } from './tavily.js';
+//# sourceMappingURL=index.js.map

package/node_modules/@noodle-borg/knowledge-crawl/dist/index.js.map

@@ -0,0 +1 @@
+{"version":3,"file":"index.js","sourceRoot":"","sources":["../src/index.ts"],"names":[],"mappings":"AAAA,OAAO,EAAE,oBAAoB,EAAoC,MAAM,gBAAgB,CAAC;AACxF,OAAO,EAAE,qBAAqB,EAAqC,MAAM,kBAAkB,CAAC;AAE5F,OAAO,EAAE,iBAAiB,EAAiC,MAAM,aAAa,CAAC"}

package/node_modules/@noodle-borg/knowledge-crawl/dist/origin.d.ts

@@ -0,0 +1,3 @@
+export declare function parseOrigin(origin: string): URL;
+export declare function safeUrl(href: string, base?: string): URL | undefined;
+//# sourceMappingURL=origin.d.ts.map

package/node_modules/@noodle-borg/knowledge-crawl/dist/origin.d.ts.map

@@ -0,0 +1 @@
+{"version":3,"file":"origin.d.ts","sourceRoot":"","sources":["../src/origin.ts"],"names":[],"mappings":"AAIA,wBAAgB,WAAW,CAAC,MAAM,EAAE,MAAM,GAAG,GAAG,CAe/C;AAED,wBAAgB,OAAO,CAAC,IAAI,EAAE,MAAM,EAAE,IAAI,CAAC,EAAE,MAAM,GAAG,GAAG,GAAG,SAAS,CAOpE"}

package/node_modules/@noodle-borg/knowledge-crawl/dist/origin.js

@@ -0,0 +1,30 @@
+/** Shared origin/URL validation for every fetcher (managed and BYO). */
+import { isIP } from 'node:net';
+import { KnowledgeError } from '@noodle-borg/knowledge';
+export function parseOrigin(origin) {
+    let url;
+    try {
+        url = new URL(origin);
+    }
+    catch {
+        throw new KnowledgeError('request', `site origin ${origin} is not a valid URL`);
+    }
+    if (url.protocol !== 'https:') {
+        throw new KnowledgeError('request', 'site origin must use https');
+    }
+    if (isIP(url.hostname.replace(/^\[|\]$/g, '')) !== 0) {
+        // A literal-IP host has no DNS step, which would bypass the socket-level SSRF guard.
+        throw new KnowledgeError('request', 'site origin must be a DNS hostname, not an IP address');
+    }
+    return url;
+}
+export function safeUrl(href, base) {
+    try {
+        const url = base === undefined ? new URL(href) : new URL(href, base);
+        return url.protocol === 'https:' ? url : undefined;
+    }
+    catch {
+        return undefined;
+    }
+}
+//# sourceMappingURL=origin.js.map

package/node_modules/@noodle-borg/knowledge-crawl/dist/origin.js.map

@@ -0,0 +1 @@
+{"version":3,"file":"origin.js","sourceRoot":"","sources":["../src/origin.ts"],"names":[],"mappings":"AAAA,wEAAwE;AACxE,OAAO,EAAE,IAAI,EAAE,MAAM,UAAU,CAAC;AAChC,OAAO,EAAE,cAAc,EAAE,MAAM,wBAAwB,CAAC;AAExD,MAAM,UAAU,WAAW,CAAC,MAAc;IACxC,IAAI,GAAQ,CAAC;IACb,IAAI,CAAC;QACH,GAAG,GAAG,IAAI,GAAG,CAAC,MAAM,CAAC,CAAC;IACxB,CAAC;IAAC,MAAM,CAAC;QACP,MAAM,IAAI,cAAc,CAAC,SAAS,EAAE,eAAe,MAAM,qBAAqB,CAAC,CAAC;IAClF,CAAC;IACD,IAAI,GAAG,CAAC,QAAQ,KAAK,QAAQ,EAAE,CAAC;QAC9B,MAAM,IAAI,cAAc,CAAC,SAAS,EAAE,4BAA4B,CAAC,CAAC;IACpE,CAAC;IACD,IAAI,IAAI,CAAC,GAAG,CAAC,QAAQ,CAAC,OAAO,CAAC,UAAU,EAAE,EAAE,CAAC,CAAC,KAAK,CAAC,EAAE,CAAC;QACrD,qFAAqF;QACrF,MAAM,IAAI,cAAc,CAAC,SAAS,EAAE,uDAAuD,CAAC,CAAC;IAC/F,CAAC;IACD,OAAO,GAAG,CAAC;AACb,CAAC;AAED,MAAM,UAAU,OAAO,CAAC,IAAY,EAAE,IAAa;IACjD,IAAI,CAAC;QACH,MAAM,GAAG,GAAG,IAAI,KAAK,SAAS,CAAC,CAAC,CAAC,IAAI,GAAG,CAAC,IAAI,CAAC,CAAC,CAAC,CAAC,IAAI,GAAG,CAAC,IAAI,EAAE,IAAI,CAAC,CAAC;QACrE,OAAO,GAAG,CAAC,QAAQ,KAAK,QAAQ,CAAC,CAAC,CAAC,GAAG,CAAC,CAAC,CAAC,SAAS,CAAC;IACrD,CAAC;IAAC,MAAM,CAAC;QACP,OAAO,SAAS,CAAC;IACnB,CAAC;AACH,CAAC"}

package/node_modules/@noodle-borg/knowledge-crawl/dist/ports.d.ts

@@ -0,0 +1,28 @@
+/**
+ * Provider-neutral site crawling port (ADR 0202 amendment 2026-08-18): one bounded call in, a
+ * deterministic ordered list of plain-text pages out. Every fetcher — the managed first-party
+ * crawler and the BYO provider adapters — satisfies this contract and its shared tests; provider
+ * concepts (job ids, credits, raw payloads) never cross it.
+ */
+export interface SiteCrawlRequest {
+    /** Exact HTTPS origin, no path. Literal-IP hosts are rejected (they bypass the DNS SSRF guard). */
+    readonly origin: string;
+    /** Path globs a page must match to be indexed; discovery may touch only these plus the root. */
+    readonly include: readonly string[];
+    readonly maxPages: number;
+    readonly maxPageBytes: number;
+}
+export interface CrawlPage {
+    readonly url: string;
+    readonly title: string;
+    readonly text: string;
+}
+export interface SiteFetcher {
+    /**
+     * Crawl the site within the request bounds. Returns pages sorted by URL; individual page
+     * failures are skipped, a total provider failure throws a `KnowledgeError` with an
+     * attributable layer and no content.
+     */
+    fetchSite(request: SiteCrawlRequest): Promise<readonly CrawlPage[]>;
+}
+//# sourceMappingURL=ports.d.ts.map

package/node_modules/@noodle-borg/knowledge-crawl/dist/ports.d.ts.map

@@ -0,0 +1 @@
+{"version":3,"file":"ports.d.ts","sourceRoot":"","sources":["../src/ports.ts"],"names":[],"mappings":"AAAA;;;;;GAKG;AAEH,MAAM,WAAW,gBAAgB;IAC/B,mGAAmG;IACnG,QAAQ,CAAC,MAAM,EAAE,MAAM,CAAC;IACxB,gGAAgG;IAChG,QAAQ,CAAC,OAAO,EAAE,SAAS,MAAM,EAAE,CAAC;IACpC,QAAQ,CAAC,QAAQ,EAAE,MAAM,CAAC;IAC1B,QAAQ,CAAC,YAAY,EAAE,MAAM,CAAC;CAC/B;AAED,MAAM,WAAW,SAAS;IACxB,QAAQ,CAAC,GAAG,EAAE,MAAM,CAAC;IACrB,QAAQ,CAAC,KAAK,EAAE,MAAM,CAAC;IACvB,QAAQ,CAAC,IAAI,EAAE,MAAM,CAAC;CACvB;AAED,MAAM,WAAW,WAAW;IAC1B;;;;OAIG;IACH,SAAS,CAAC,OAAO,EAAE,gBAAgB,GAAG,OAAO,CAAC,SAAS,SAAS,EAAE,CAAC,CAAC;CACrE"}

package/node_modules/@noodle-borg/knowledge-crawl/dist/ports.js

@@ -0,0 +1,8 @@
+/**
+ * Provider-neutral site crawling port (ADR 0202 amendment 2026-08-18): one bounded call in, a
+ * deterministic ordered list of plain-text pages out. Every fetcher — the managed first-party
+ * crawler and the BYO provider adapters — satisfies this contract and its shared tests; provider
+ * concepts (job ids, credits, raw payloads) never cross it.
+ */
+export {};
+//# sourceMappingURL=ports.js.map

package/node_modules/@noodle-borg/knowledge-crawl/dist/ports.js.map

@@ -0,0 +1 @@
+{"version":3,"file":"ports.js","sourceRoot":"","sources":["../src/ports.ts"],"names":[],"mappings":"AAAA;;;;;GAKG"}

package/node_modules/@noodle-borg/knowledge-crawl/dist/robots.d.ts

@@ -0,0 +1,12 @@
+/**
+ * Minimal robots.txt evaluation for the managed crawler: `*` user-agent groups only, with the
+ * de-facto longest-match precedence between Allow and Disallow (Allow wins an exact-length tie),
+ * plus global `Sitemap:` collection. Deliberately dependency-free; if richer semantics are ever
+ * needed, verify the dependency license first (root rules).
+ */
+export interface Robots {
+    isAllowed(path: string): boolean;
+    readonly sitemaps: readonly string[];
+}
+export declare function parseRobots(text: string): Robots;
+//# sourceMappingURL=robots.d.ts.map

package/node_modules/@noodle-borg/knowledge-crawl/dist/robots.d.ts.map

@@ -0,0 +1 @@
+{"version":3,"file":"robots.d.ts","sourceRoot":"","sources":["../src/robots.ts"],"names":[],"mappings":"AAAA;;;;;GAKG;AAOH,MAAM,WAAW,MAAM;IACrB,SAAS,CAAC,IAAI,EAAE,MAAM,GAAG,OAAO,CAAC;IACjC,QAAQ,CAAC,QAAQ,EAAE,SAAS,MAAM,EAAE,CAAC;CACtC;AAED,wBAAgB,WAAW,CAAC,IAAI,EAAE,MAAM,GAAG,MAAM,CAmDhD"}

package/node_modules/@noodle-borg/knowledge-crawl/dist/robots.js

@@ -0,0 +1,64 @@
+/**
+ * Minimal robots.txt evaluation for the managed crawler: `*` user-agent groups only, with the
+ * de-facto longest-match precedence between Allow and Disallow (Allow wins an exact-length tie),
+ * plus global `Sitemap:` collection. Deliberately dependency-free; if richer semantics are ever
+ * needed, verify the dependency license first (root rules).
+ */
+export function parseRobots(text) {
+    const rules = [];
+    const sitemaps = [];
+    let inStarGroup = false;
+    let groupHasAgents = false;
+    for (const rawLine of text.split(/\r?\n/)) {
+        const line = rawLine.replace(/#.*$/, '').trim();
+        if (line === '')
+            continue;
+        const separator = line.indexOf(':');
+        if (separator === -1)
+            continue;
+        const field = line.slice(0, separator).trim().toLowerCase();
+        const value = line.slice(separator + 1).trim();
+        if (field === 'sitemap') {
+            if (value !== '')
+                sitemaps.push(value);
+            continue;
+        }
+        if (field === 'user-agent') {
+            // Consecutive user-agent lines share one group; a rule line closes the agent list, so a
+            // user-agent line after rules starts a fresh group.
+            if (groupHasAgents) {
+                inStarGroup = false;
+                groupHasAgents = false;
+            }
+            if (value === '*')
+                inStarGroup = true;
+            continue;
+        }
+        if (field === 'allow' || field === 'disallow') {
+            groupHasAgents = true;
+            if (!inStarGroup)
+                continue;
+            // A bare "Disallow:" means allow-all per the de-facto standard.
+            if (value === '')
+                continue;
+            rules.push({ allow: field === 'allow', path: value });
+        }
+    }
+    return {
+        sitemaps,
+        isAllowed(path) {
+            let best;
+            for (const rule of rules) {
+                if (!path.startsWith(rule.path))
+                    continue;
+                if (best === undefined ||
+                    rule.path.length > best.path.length ||
+                    (rule.path.length === best.path.length && rule.allow)) {
+                    best = rule;
+                }
+            }
+            return best?.allow ?? true;
+        },
+    };
+}
+//# sourceMappingURL=robots.js.map

package/node_modules/@noodle-borg/knowledge-crawl/dist/robots.js.map

@@ -0,0 +1 @@
+{"version":3,"file":"robots.js","sourceRoot":"","sources":["../src/robots.ts"],"names":[],"mappings":"AAAA;;;;;GAKG;AAYH,MAAM,UAAU,WAAW,CAAC,IAAY;IACtC,MAAM,KAAK,GAAiB,EAAE,CAAC;IAC/B,MAAM,QAAQ,GAAa,EAAE,CAAC;IAC9B,IAAI,WAAW,GAAG,KAAK,CAAC;IACxB,IAAI,cAAc,GAAG,KAAK,CAAC;IAC3B,KAAK,MAAM,OAAO,IAAI,IAAI,CAAC,KAAK,CAAC,OAAO,CAAC,EAAE,CAAC;QAC1C,MAAM,IAAI,GAAG,OAAO,CAAC,OAAO,CAAC,MAAM,EAAE,EAAE,CAAC,CAAC,IAAI,EAAE,CAAC;QAChD,IAAI,IAAI,KAAK,EAAE;YAAE,SAAS;QAC1B,MAAM,SAAS,GAAG,IAAI,CAAC,OAAO,CAAC,GAAG,CAAC,CAAC;QACpC,IAAI,SAAS,KAAK,CAAC,CAAC;YAAE,SAAS;QAC/B,MAAM,KAAK,GAAG,IAAI,CAAC,KAAK,CAAC,CAAC,EAAE,SAAS,CAAC,CAAC,IAAI,EAAE,CAAC,WAAW,EAAE,CAAC;QAC5D,MAAM,KAAK,GAAG,IAAI,CAAC,KAAK,CAAC,SAAS,GAAG,CAAC,CAAC,CAAC,IAAI,EAAE,CAAC;QAC/C,IAAI,KAAK,KAAK,SAAS,EAAE,CAAC;YACxB,IAAI,KAAK,KAAK,EAAE;gBAAE,QAAQ,CAAC,IAAI,CAAC,KAAK,CAAC,CAAC;YACvC,SAAS;QACX,CAAC;QACD,IAAI,KAAK,KAAK,YAAY,EAAE,CAAC;YAC3B,wFAAwF;YACxF,oDAAoD;YACpD,IAAI,cAAc,EAAE,CAAC;gBACnB,WAAW,GAAG,KAAK,CAAC;gBACpB,cAAc,GAAG,KAAK,CAAC;YACzB,CAAC;YACD,IAAI,KAAK,KAAK,GAAG;gBAAE,WAAW,GAAG,IAAI,CAAC;YACtC,SAAS;QACX,CAAC;QACD,IAAI,KAAK,KAAK,OAAO,IAAI,KAAK,KAAK,UAAU,EAAE,CAAC;YAC9C,cAAc,GAAG,IAAI,CAAC;YACtB,IAAI,CAAC,WAAW;gBAAE,SAAS;YAC3B,gEAAgE;YAChE,IAAI,KAAK,KAAK,EAAE;gBAAE,SAAS;YAC3B,KAAK,CAAC,IAAI,CAAC,EAAE,KAAK,EAAE,KAAK,KAAK,OAAO,EAAE,IAAI,EAAE,KAAK,EAAE,CAAC,CAAC;QACxD,CAAC;IACH,CAAC;IACD,OAAO;QACL,QAAQ;QACR,SAAS,CAAC,IAAY;YACpB,IAAI,IAA4B,CAAC;YACjC,KAAK,MAAM,IAAI,IAAI,KAAK,EAAE,CAAC;gBACzB,IAAI,CAAC,IAAI,CAAC,UAAU,CAAC,IAAI,CAAC,IAAI,CAAC;oBAAE,SAAS;gBAC1C,IACE,IAAI,KAAK,SAAS;oBAClB,IAAI,CAAC,IAAI,CAAC,MAAM,GAAG,IAAI,CAAC,IAAI,CAAC,MAAM;oBACnC,CAAC,IAAI,CAAC,IAAI,CAAC,MAAM,KAAK,IAAI,CAAC,IAAI,CAAC,MAAM,IAAI,IAAI,CAAC,KAAK,CAAC,EACrD,CAAC;oBACD,IAAI,GAAG,IAAI,CAAC;gBACd,CAAC;YACH,CAAC;YACD,OAAO,IAAI,EAAE,KAAK,IAAI,IAAI,CAAC;QAC7B,CAAC;KACF,CAAC;AACJ,CAAC"}

package/node_modules/@noodle-borg/knowledge-crawl/dist/sitemap.d.ts

@@ -0,0 +1,3 @@
+/** Bounded `<loc>` extraction from sitemap XML — locations only, nothing else is read. */
+export declare function parseSitemapLocations(xml: string): readonly string[];
+//# sourceMappingURL=sitemap.d.ts.map

package/node_modules/@noodle-borg/knowledge-crawl/dist/sitemap.d.ts.map

@@ -0,0 +1 @@
+{"version":3,"file":"sitemap.d.ts","sourceRoot":"","sources":["../src/sitemap.ts"],"names":[],"mappings":"AAAA,0FAA0F;AAC1F,wBAAgB,qBAAqB,CAAC,GAAG,EAAE,MAAM,GAAG,SAAS,MAAM,EAAE,CAOpE"}

package/node_modules/@noodle-borg/knowledge-crawl/dist/sitemap.js

@@ -0,0 +1,11 @@
+/** Bounded `<loc>` extraction from sitemap XML — locations only, nothing else is read. */
+export function parseSitemapLocations(xml) {
+    const locations = [];
+    for (const match of xml.matchAll(/<loc>\s*([^<]+?)\s*<\/loc>/gi)) {
+        const location = match[1];
+        if (location !== undefined && location !== '')
+            locations.push(location);
+    }
+    return locations;
+}
+//# sourceMappingURL=sitemap.js.map

package/node_modules/@noodle-borg/knowledge-crawl/dist/sitemap.js.map

@@ -0,0 +1 @@
+{"version":3,"file":"sitemap.js","sourceRoot":"","sources":["../src/sitemap.ts"],"names":[],"mappings":"AAAA,0FAA0F;AAC1F,MAAM,UAAU,qBAAqB,CAAC,GAAW;IAC/C,MAAM,SAAS,GAAa,EAAE,CAAC;IAC/B,KAAK,MAAM,KAAK,IAAI,GAAG,CAAC,QAAQ,CAAC,8BAA8B,CAAC,EAAE,CAAC;QACjE,MAAM,QAAQ,GAAG,KAAK,CAAC,CAAC,CAAC,CAAC;QAC1B,IAAI,QAAQ,KAAK,SAAS,IAAI,QAAQ,KAAK,EAAE;YAAE,SAAS,CAAC,IAAI,CAAC,QAAQ,CAAC,CAAC;IAC1E,CAAC;IACD,OAAO,SAAS,CAAC;AACnB,CAAC"}

package/node_modules/@noodle-borg/knowledge-crawl/dist/tavily.d.ts

@@ -0,0 +1,11 @@
+import type { CrawlPage, SiteCrawlRequest, SiteFetcher } from './ports.js';
+export interface TavilySiteFetcherOptions {
+    readonly apiKey: string;
+    readonly fetchImpl?: typeof fetch;
+}
+export declare class TavilySiteFetcher implements SiteFetcher {
+    #private;
+    constructor(options: TavilySiteFetcherOptions);
+    fetchSite(request: SiteCrawlRequest): Promise<readonly CrawlPage[]>;
+}
+//# sourceMappingURL=tavily.d.ts.map

package/node_modules/@noodle-borg/knowledge-crawl/dist/tavily.d.ts.map

@@ -0,0 +1 @@
+{"version":3,"file":"tavily.d.ts","sourceRoot":"","sources":["../src/tavily.ts"],"names":[],"mappings":"AAWA,OAAO,KAAK,EAAE,SAAS,EAAE,gBAAgB,EAAE,WAAW,EAAE,MAAM,YAAY,CAAC;AAI3E,MAAM,WAAW,wBAAwB;IACvC,QAAQ,CAAC,MAAM,EAAE,MAAM,CAAC;IACxB,QAAQ,CAAC,SAAS,CAAC,EAAE,OAAO,KAAK,CAAC;CACnC;AAED,qBAAa,iBAAkB,YAAW,WAAW;;gBAIvC,OAAO,EAAE,wBAAwB;IAKvC,SAAS,CAAC,OAAO,EAAE,gBAAgB,GAAG,OAAO,CAAC,SAAS,SAAS,EAAE,CAAC;CAqC1E"}

package/node_modules/@noodle-borg/knowledge-crawl/dist/tavily.js

@@ -0,0 +1,53 @@
+/**
+ * BYO Tavily fetcher (ADR 0202 amendment 2026-08-18): the customer's own Tavily account crawls
+ * the site through the synchronous hosted REST API. Wire shape transcribed from the official API
+ * reference 2026-08-18: `POST https://api.tavily.com/crawl` with Bearer auth and
+ * `{ url, limit, allow_external, format, extract_depth }` → `{ results: [{ url, raw_content }] }`.
+ * Crawled pages carry no separate title field, so the URL stands in. Errors are attributable and
+ * content-free — provider bodies and the API key never cross the port.
+ */
+import { KnowledgeError } from '@noodle-borg/knowledge';
+import { boundPages } from './bounds.js';
+import { parseOrigin } from './origin.js';
+const API = 'https://api.tavily.com/crawl';
+export class TavilySiteFetcher {
+    #apiKey;
+    #fetch;
+    constructor(options) {
+        this.#apiKey = options.apiKey;
+        this.#fetch = options.fetchImpl ?? fetch;
+    }
+    async fetchSite(request) {
+        const origin = parseOrigin(request.origin);
+        let response;
+        try {
+            response = await this.#fetch(API, {
+                method: 'POST',
+                headers: {
+                    authorization: `Bearer ${this.#apiKey}`,
+                    'content-type': 'application/json',
+                },
+                body: JSON.stringify({
+                    url: origin.origin,
+                    limit: request.maxPages,
+                    allow_external: false,
+                    format: 'text',
+                    extract_depth: 'basic',
+                }),
+            });
+        }
+        catch {
+            throw new KnowledgeError('provider', 'tavily crawl request failed');
+        }
+        if (!response.ok) {
+            throw new KnowledgeError('provider', `tavily crawl returned HTTP ${response.status}`);
+        }
+        const body = (await response.json().catch(() => ({})));
+        return boundPages((body.results ?? []).map((result) => ({
+            url: result.url ?? '',
+            title: '',
+            text: result.raw_content ?? '',
+        })), request, origin);
+    }
+}
+//# sourceMappingURL=tavily.js.map

package/node_modules/@noodle-borg/knowledge-crawl/dist/tavily.js.map

@@ -0,0 +1 @@
+{"version":3,"file":"tavily.js","sourceRoot":"","sources":["../src/tavily.ts"],"names":[],"mappings":"AAAA;;;;;;;GAOG;AACH,OAAO,EAAE,cAAc,EAAE,MAAM,wBAAwB,CAAC;AACxD,OAAO,EAAE,UAAU,EAAE,MAAM,aAAa,CAAC;AACzC,OAAO,EAAE,WAAW,EAAE,MAAM,aAAa,CAAC;AAG1C,MAAM,GAAG,GAAG,8BAA8B,CAAC;AAO3C,MAAM,OAAO,iBAAiB;IACnB,OAAO,CAAS;IAChB,MAAM,CAAe;IAE9B,YAAY,OAAiC;QAC3C,IAAI,CAAC,OAAO,GAAG,OAAO,CAAC,MAAM,CAAC;QAC9B,IAAI,CAAC,MAAM,GAAG,OAAO,CAAC,SAAS,IAAI,KAAK,CAAC;IAC3C,CAAC;IAED,KAAK,CAAC,SAAS,CAAC,OAAyB;QACvC,MAAM,MAAM,GAAG,WAAW,CAAC,OAAO,CAAC,MAAM,CAAC,CAAC;QAC3C,IAAI,QAAkB,CAAC;QACvB,IAAI,CAAC;YACH,QAAQ,GAAG,MAAM,IAAI,CAAC,MAAM,CAAC,GAAG,EAAE;gBAChC,MAAM,EAAE,MAAM;gBACd,OAAO,EAAE;oBACP,aAAa,EAAE,UAAU,IAAI,CAAC,OAAO,EAAE;oBACvC,cAAc,EAAE,kBAAkB;iBACnC;gBACD,IAAI,EAAE,IAAI,CAAC,SAAS,CAAC;oBACnB,GAAG,EAAE,MAAM,CAAC,MAAM;oBAClB,KAAK,EAAE,OAAO,CAAC,QAAQ;oBACvB,cAAc,EAAE,KAAK;oBACrB,MAAM,EAAE,MAAM;oBACd,aAAa,EAAE,OAAO;iBACvB,CAAC;aACH,CAAC,CAAC;QACL,CAAC;QAAC,MAAM,CAAC;YACP,MAAM,IAAI,cAAc,CAAC,UAAU,EAAE,6BAA6B,CAAC,CAAC;QACtE,CAAC;QACD,IAAI,CAAC,QAAQ,CAAC,EAAE,EAAE,CAAC;YACjB,MAAM,IAAI,cAAc,CAAC,UAAU,EAAE,8BAA8B,QAAQ,CAAC,MAAM,EAAE,CAAC,CAAC;QACxF,CAAC;QACD,MAAM,IAAI,GAAG,CAAC,MAAM,QAAQ,CAAC,IAAI,EAAE,CAAC,KAAK,CAAC,GAAG,EAAE,CAAC,CAAC,EAAE,CAAC,CAAC,CAEpD,CAAC;QACF,OAAO,UAAU,CACf,CAAC,IAAI,CAAC,OAAO,IAAI,EAAE,CAAC,CAAC,GAAG,CAAC,CAAC,MAAM,EAAE,EAAE,CAAC,CAAC;YACpC,GAAG,EAAE,MAAM,CAAC,GAAG,IAAI,EAAE;YACrB,KAAK,EAAE,EAAE;YACT,IAAI,EAAE,MAAM,CAAC,WAAW,IAAI,EAAE;SAC/B,CAAC,CAAC,EACH,OAAO,EACP,MAAM,CACP,CAAC;IACJ,CAAC;CACF"}

package/node_modules/@noodle-borg/knowledge-crawl/package.json

@@ -0,0 +1,32 @@
+{
+  "name": "@noodle-borg/knowledge-crawl",
+  "version": "0.0.0",
+  "private": true,
+  "license": "Apache-2.0",
+  "type": "module",
+  "engines": {
+    "node": ">=24"
+  },
+  "main": "./dist/index.js",
+  "types": "./dist/index.d.ts",
+  "exports": {
+    ".": {
+      "types": "./dist/index.d.ts",
+      "import": "./dist/index.js"
+    }
+  },
+  "files": [
+    "dist"
+  ],
+  "scripts": {
+    "build": "tsc -p tsconfig.json",
+    "typecheck": "tsc -p tsconfig.json --noEmit"
+  },
+  "dependencies": {
+    "@noodle-borg/connector-http": "0.0.0",
+    "@noodle-borg/knowledge": "0.0.0"
+  },
+  "devDependencies": {
+    "@types/node": "^24.13.1"
+  }
+}