sbuilder-mcp 0.20.0 → 0.21.0

package/CHANGELOG.md

@@ -6,6 +6,14 @@ All notable changes to this project are documented in this file.
 The format is based on [Keep a Changelog](https://keepachangelog.com/en/1.1.0/),
 and this project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0.html).
 
+## [0.21.0] - 2026-09-10
+
+### Added
+- sb_import and sb_import_site now detect a `<form>` on the source page and report it as `forms_found` (field count and labels) instead of silently dropping it, pointing the caller at `sb_store action:"form"` to rebuild it with a valid field vocabulary.
+
+### Fixed
+- sb_import_site now rewrites links between the pages it imports so the new site's menu points at itself instead of back at the source it was copied from, and reports any same-origin links left pointing off-site under `links.still_off_site` so the caller knows to raise `max_pages`.
+
 ## [0.20.0] - 2026-09-10
 
 ### Added

package/CHANGELOG.vi.md

@@ -6,6 +6,14 @@ Mọi thay đổi đáng chú ý của dự án được ghi lại trong file n
 Định dạng dựa trên [Keep a Changelog](https://keepachangelog.com/en/1.1.0/),
 và dự án tuân theo [Semantic Versioning](https://semver.org/spec/v2.0.0.html).
 
+## [0.21.0] - 2026-09-10
+
+### Added
+- sb_import và sb_import_site giờ phát hiện `<form>` trên trang nguồn và báo cáo qua `forms_found` (số field và nhãn) thay vì âm thầm bỏ qua, đồng thời chỉ người gọi tới `sb_store action:"form"` để dựng lại form với vocabulary field hợp lệ.
+
+### Fixed
+- sb_import_site giờ viết lại các link giữa các trang mà nó import, để menu của site mới trỏ về chính nó thay vì trỏ ngược về site nguồn đã copy, và báo cáo các link cùng origin còn trỏ ra ngoài qua `links.still_off_site` để người gọi biết cần tăng `max_pages`.
+
 ## [0.20.0] - 2026-09-10
 
 ### Added

package/dist/domains/site/importmap.js

@@ -1,5 +1,6 @@
 import { walk } from '../../core/tree.js';
 import { stickySeeds } from './sticky.js';
+import { normalizeUrl } from './discover.js';
 import { ICON_NAMES } from '../../catalog/icons.generated.js';
 /** Style keys read off a node, ignoring anything unset. */
 function styleOf(n) {
@@ -416,3 +417,39 @@ export function rehostImages(captured, map) {
     });
     return captured.map(visit);
 }
+/**
+ * POINT THE IMPORTED LINKS AT THE IMPORTED PAGES.
+ *
+ * A captured link keeps the SOURCE's absolute URL, so a site brought over with
+ * `sb_import_site` had a menu that sent every visitor back to the website it was
+ * copied from — twelve pages built here and not one way to reach any of them.
+ * The most basic feature a website has, and the import was quietly working
+ * against it.
+ *
+ * Only the links whose target was ACTUALLY IMPORTED are rewritten. A same-origin
+ * link to a page the cap left out is counted rather than pointed at a slug that
+ * does not exist here: an off-site link that works beats a local one that 404s,
+ * and the count is what tells the caller to raise `max_pages`. A genuinely
+ * external link is left alone and is not interesting.
+ */
+export function relink(captured, local, origin) {
+    let rewritten = 0;
+    let unimported = 0;
+    const one = (c) => {
+        let href = c.href;
+        if (href) {
+            const norm = normalizeUrl(href);
+            const to = norm ? local.get(norm) : undefined;
+            if (to) {
+                href = to;
+                rewritten += 1;
+            }
+            else if (norm && norm.indexOf(origin) === 0) {
+                unimported += 1;
+            }
+        }
+        const kids = c.children ? c.children.map(one) : undefined;
+        return { ...c, ...(href ? { href } : {}), ...(kids ? { children: kids } : {}) };
+    };
+    return { sections: captured.map(one), rewritten, unimported };
+}

package/dist/tools/importpage.js

@@ -4,7 +4,7 @@ import { capture, captureMany, crawlLinks } from '../vision/capture.js';
 import { uploadMedia } from '../transport/media.js';
 import { addSubtree } from '../domains/site/builder.js';
 import { middleEnd } from '../domains/site/traps.js';
-import { toSpecs, tokensFromPage, imageSources, rehostImages, } from '../domains/site/importmap.js';
+import { toSpecs, tokensFromPage, imageSources, rehostImages, relink, } from '../domains/site/importmap.js';
 import { canonFor, choosePages, normalizeUrl, robotsRules, robotsSitemaps, sitemapUrls, } from '../domains/site/discover.js';
 import { loadSource } from '../transport/pages.js';
 import { PageDoc } from '../domains/site/document.js';
@@ -207,6 +207,7 @@ export function registerImportTools(server, ctx, session) {
                 title: shot.title,
                 sections: specs.length,
                 images: images.length,
+                ...(shot.forms?.length ? { forms_found: shot.forms } : {}),
                 tokens,
                 skipped: shot.skipped,
                 note: 'Structure and content only — the source\'s CSS and layout are NOT copied, and the ' +
@@ -278,6 +279,7 @@ export function registerImportTools(server, ctx, session) {
         return text({
             read: shot.url,
             added_sections: added,
+            ...(shot.forms?.length ? { forms_found: shot.forms } : {}),
             // WHAT WAS LEFT BEHIND, on the real run too. The dry run said it and the
             // real one did not, which is the wrong way round: a caller who skipped
             // the preview is exactly the caller who needs to be told that 21 nodes
@@ -477,6 +479,16 @@ export function registerImportTools(server, ctx, session) {
         // cannot be — each page is its own create and its own save — so the honest
         // shape is per-page outcomes. Aborting on the fourth of twelve would leave
         // three pages built, nine not, and no report saying which.
+        // WHERE EACH IMPORTED PAGE WILL LIVE HERE, decided before the first one is
+        // built — page two's link to page seven has to work, and page seven does
+        // not exist yet. The entry resolves to "/" when it merges into the site's
+        // own home page, because that is the address it will answer on.
+        const origin = new URL(entry).origin;
+        const localPath = new Map();
+        for (const p of plan.pages) {
+            const isEntry = p.url === entry;
+            localPath.set(p.url, isEntry && homepage !== false && home ? '/' : `/${p.slug}`);
+        }
         const built = [];
         const failed = [];
         // WHAT EACH PAGE SAYS ITS OWN ADDRESS IS. A sitemap cannot tell you that
@@ -486,6 +498,9 @@ export function registerImportTools(server, ctx, session) {
         // under two slugs, and nothing in the plan looks wrong.
         const identities = new Set();
         const aliased = [];
+        let relinked = 0;
+        let unimported = 0;
+        const formsSeen = [];
         let lastOpened = '';
         for (const p of plan.pages) {
             const shot = byUrl.get(p.url);
@@ -500,8 +515,13 @@ export function registerImportTools(server, ctx, session) {
             }
             identities.add(identity);
             try {
-                const sections = rehosted.size > 0 ? rehostImages(shot.result.sections, rehosted) : shot.result.sections;
-                const specs = toSpecs(sections, tokens);
+                const hosted = rehosted.size > 0 ? rehostImages(shot.result.sections, rehosted) : shot.result.sections;
+                for (const f of shot.result.forms ?? [])
+                    formsSeen.push({ page: p.slug, ...f });
+                const linked = relink(hosted, localPath, origin);
+                relinked += linked.rewritten;
+                unimported += linked.unimported;
+                const specs = toSpecs(linked.sections, tokens);
                 if (specs.length === 0) {
                     failed.push({
                         url: p.url,
@@ -583,6 +603,19 @@ export function registerImportTools(server, ctx, session) {
             built,
             ...(failed.length ? { failed } : {}),
             ...(aliased.length ? { same_page: aliased } : {}),
+            links: {
+                rewritten: relinked,
+                ...(unimported ? { still_off_site: unimported } : {}),
+            },
+            ...(formsSeen.length
+                ? {
+                    forms_found: formsSeen,
+                    forms_note: 'A form is NOT a page node here — it lives in its own document, and its fields ' +
+                        'are a vocabulary the server validates. sb_store action:"form" seeds any of the ' +
+                        "platform's 17 templates (contact, subscribe, booking, login, register …) with " +
+                        'that document already correct; place the form on the page after.',
+                }
+                : {}),
             ...(Object.keys(plan.skipped).length || found.aliases
                 ? { skipped: { ...plan.skipped, ...(found.aliases ? { 'canonical-alias': found.aliases } : {}) } }
                 : {}),

package/dist/vision/capture.js

@@ -70,8 +70,13 @@ function capturePage(limits) {
     // one kind of content that could not survive the trip at all.
     const IGNORE = new Set([
         'SCRIPT', 'STYLE', 'NOSCRIPT', 'TEMPLATE', 'SVG', 'PATH', 'CANVAS',
-        'NAV', 'FORM', 'INPUT', 'SELECT', 'TEXTAREA', 'BUTTON',
+        // FORM is NOT here: its own branch below records it before returning
+        // nothing, so a contact page says it had one. The CONTROLS stay ignored —
+        // a stray input outside a form is chrome, and the fields of a form that IS
+        // reported are counted there rather than walked into.
+        'NAV', 'INPUT', 'SELECT', 'TEXTAREA', 'BUTTON',
     ]);
+    const forms = [];
     /** The provider and id behind an embed URL, or null if this platform has no element for it. */
     const embedOf = (raw) => {
         const u = raw.split('?')[0];
@@ -322,6 +327,23 @@ function capturePage(limits) {
                 return [{ kind: 'accordion', children: [{ kind: 'accordion-item', text: label, children: body }] }];
             }
             // A RULE BETWEEN SECTIONS IS A DESIGN DECISION, and it is one node.
+            // A FORM IS SEEN AND NOT REBUILT. Its fields are a vocabulary the server
+            // validates (`mapTo`), and `sb_store action:"form"` owns that; what this
+            // walk can honestly do is say the page had one, so a contact page does not
+            // arrive with no way to contact anybody and nothing saying why.
+            if (tag === 'FORM') {
+                const controls = Array.from(el.querySelectorAll('input, textarea, select'));
+                const labels = [];
+                for (const l of Array.from(el.querySelectorAll('label'))) {
+                    const t = clean(l.textContent);
+                    if (t && labels.length < 10)
+                        labels.push(t);
+                }
+                if (controls.length > 0)
+                    forms.push({ fields: controls.length, labels });
+                skip('form');
+                return [];
+            }
             if (tag === 'HR') {
                 taken.nodes++;
                 return [{ kind: 'divider' }];
@@ -608,6 +630,7 @@ function capturePage(limits) {
         url: here,
         title: clean(document.title),
         ...(canonical ? { canonical: abs(canonical) } : {}),
+        ...(forms.length ? { forms } : {}),
         sections,
         skipped,
     };

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "sbuilder-mcp",
-  "version": "0.20.0",
+  "version": "0.21.0",
   "description": "MCP server that designs and operates a Store Builder site — pages, data, theme and publish — through the platform's own API and live-edit protocol.",
   "mcpName": "io.github.vuluu2k/sbuilder-mcp",
   "type": "module",