@mgsoftwarebv/mg-dashboard-mcp 7.1.0 → 7.3.0

package/dist/index.js

@@ -956,9 +956,9 @@ async function fetchAndFormatRun(conn, proxy, sshExec2, instance, runId) {
     return { content: [{ type: "text", text: `Invalid API response:
 ${rawJson.substring(0, 500)}` }] };
   }
-  let text7 = formatRunDetail(run);
-  if (logs) text7 += "\n\n--- Logs ---\n" + logs;
-  return { content: [{ type: "text", text: text7 }] };
+  let text9 = formatRunDetail(run);
+  if (logs) text9 += "\n\n--- Logs ---\n" + logs;
+  return { content: [{ type: "text", text: text9 }] };
 }
 async function waitForCompletion(conn, proxy, sshExec2, instance, runId, waitSeconds) {
   const pollInterval = 3e3;
@@ -980,10 +980,10 @@ async function waitForCompletion(conn, proxy, sshExec2, instance, runId, waitSec
       continue;
     }
     if (TERMINAL_STATUSES.has(run.status)) {
-      let text7 = formatRunDetail(run);
+      let text9 = formatRunDetail(run);
       const logs = await fetchRunLogs(runId, conn, proxy, sshExec2);
-      if (logs) text7 += "\n\n--- Logs ---\n" + logs;
-      return { content: [{ type: "text", text: text7 }] };
+      if (logs) text9 += "\n\n--- Logs ---\n" + logs;
+      return { content: [{ type: "text", text: text9 }] };
     }
   }
   return {
@@ -3670,10 +3670,10 @@ var ZodObject = class _ZodObject extends ZodType {
   //   }) as any;
   //   return merged;
   // }
-  catchall(index5) {
+  catchall(index7) {
     return new _ZodObject({
       ...this._def,
-      catchall: index5
+      catchall: index7
     });
   }
   pick(mask) {
@@ -3991,9 +3991,9 @@ function mergeValues(a, b) {
       return { valid: false };
     }
     const newArray = [];
-    for (let index5 = 0; index5 < a.length; index5++) {
-      const itemA = a[index5];
-      const itemB = b[index5];
+    for (let index7 = 0; index7 < a.length; index7++) {
+      const itemA = a[index7];
+      const itemB = b[index7];
       const sharedValue = mergeValues(itemA, itemB);
       if (!sharedValue.valid) {
         return { valid: false };
@@ -4199,10 +4199,10 @@ var ZodMap = class extends ZodType {
     }
     const keyType = this._def.keyType;
     const valueType = this._def.valueType;
-    const pairs = [...ctx.data.entries()].map(([key, value], index5) => {
+    const pairs = [...ctx.data.entries()].map(([key, value], index7) => {
       return {
-        key: keyType._parse(new ParseInputLazyPath(ctx, key, ctx.path, [index5, "key"])),
-        value: valueType._parse(new ParseInputLazyPath(ctx, value, ctx.path, [index5, "value"]))
+        key: keyType._parse(new ParseInputLazyPath(ctx, key, ctx.path, [index7, "key"])),
+        value: valueType._parse(new ParseInputLazyPath(ctx, value, ctx.path, [index7, "value"]))
       };
     });
     if (ctx.common.async) {
@@ -5170,6 +5170,231 @@ pgTable("two_factor", {
   secret: text("secret").notNull(),
   backupCodes: text("backup_codes")
 });
+var contentSourceType = pgEnum("content_source_type", [
+  "own_site",
+  "client_site",
+  "external_source",
+  "competitor",
+  "news",
+  "government",
+  "supplier",
+  "unknown"
+]);
+var contentSnapshotKind = pgEnum("content_snapshot_kind", [
+  "corpus",
+  "source"
+]);
+var contentSources = pgTable(
+  "content_sources",
+  {
+    id: uuid("id").primaryKey().defaultRandom(),
+    url: text("url").notNull(),
+    canonicalUrl: text("canonical_url"),
+    domain: text("domain").notNull(),
+    sourceType: contentSourceType("source_type").notNull().default("unknown"),
+    title: text("title"),
+    author: text("author"),
+    publishedAt: timestamp("published_at", { withTimezone: true }),
+    extractionMethod: text("extraction_method"),
+    summary: text("summary"),
+    facts: jsonb("facts").$type().notNull().default([]),
+    limitations: jsonb("limitations").$type().notNull().default([]),
+    fullText: text("full_text"),
+    textHash: text("text_hash"),
+    wordCount: integer("word_count"),
+    language: text("language"),
+    projectKey: text("project_key"),
+    siteKey: text("site_key"),
+    ticketNumber: text("ticket_number"),
+    metadata: jsonb("metadata").$type().notNull().default({}),
+    fetchedAt: timestamp("fetched_at", { withTimezone: true }).notNull().defaultNow(),
+    lastSeenAt: timestamp("last_seen_at", { withTimezone: true }).notNull().defaultNow(),
+    createdAt: timestamp("created_at", { withTimezone: true }).notNull().defaultNow(),
+    updatedAt: timestamp("updated_at", { withTimezone: true }).notNull().defaultNow()
+  },
+  (table) => [
+    uniqueIndex("content_sources_canonical_url_uidx").on(table.canonicalUrl),
+    index("content_sources_domain_idx").on(table.domain),
+    index("content_sources_type_idx").on(table.sourceType),
+    index("content_sources_text_hash_idx").on(table.textHash),
+    index("content_sources_project_idx").on(table.projectKey),
+    index("content_sources_ticket_idx").on(table.ticketNumber),
+    index("content_sources_last_seen_idx").on(table.lastSeenAt.desc())
+  ]
+);
+var contentEvidencePacks = pgTable(
+  "content_evidence_packs",
+  {
+    id: uuid("id").primaryKey().defaultRandom(),
+    topic: text("topic"),
+    locale: text("locale"),
+    projectKey: text("project_key"),
+    siteKey: text("site_key"),
+    ticketNumber: text("ticket_number"),
+    generatedAt: timestamp("generated_at", { withTimezone: true }).notNull().defaultNow(),
+    sources: jsonb("sources").$type().notNull().default([]),
+    verifiedFacts: jsonb("verified_facts").$type().notNull().default([]),
+    conflicts: jsonb("conflicts").$type().notNull().default([]),
+    uniqueClaims: jsonb("unique_claims").$type().notNull().default([]),
+    openQuestions: jsonb("open_questions").$type().notNull().default([]),
+    verbatimWarnings: jsonb("verbatim_warnings").$type().notNull().default([]),
+    limitations: jsonb("limitations").$type().notNull().default([]),
+    metadata: jsonb("metadata").$type().notNull().default({}),
+    createdAt: timestamp("created_at", { withTimezone: true }).notNull().defaultNow(),
+    updatedAt: timestamp("updated_at", { withTimezone: true }).notNull().defaultNow()
+  },
+  (table) => [
+    index("content_evidence_packs_project_idx").on(table.projectKey),
+    index("content_evidence_packs_ticket_idx").on(table.ticketNumber),
+    index("content_evidence_packs_generated_idx").on(table.generatedAt.desc())
+  ]
+);
+pgTable(
+  "content_evidence_pack_source",
+  {
+    id: uuid("id").primaryKey().defaultRandom(),
+    packId: uuid("pack_id").notNull().references(() => contentEvidencePacks.id, { onDelete: "cascade" }),
+    contentSourceId: uuid("content_source_id").notNull().references(() => contentSources.id, { onDelete: "cascade" }),
+    packSourceRef: text("pack_source_ref"),
+    createdAt: timestamp("created_at", { withTimezone: true }).notNull().defaultNow()
+  },
+  (table) => [
+    uniqueIndex("content_evidence_pack_source_uidx").on(
+      table.packId,
+      table.contentSourceId
+    ),
+    index("content_evidence_pack_source_source_idx").on(table.contentSourceId)
+  ]
+);
+var contentCorpusItems = pgTable(
+  "content_corpus_items",
+  {
+    id: uuid("id").primaryKey().defaultRandom(),
+    projectKey: text("project_key").notNull(),
+    siteKey: text("site_key"),
+    url: text("url").notNull(),
+    canonicalUrl: text("canonical_url"),
+    domain: text("domain").notNull(),
+    pagePath: text("page_path"),
+    sourceType: contentSourceType("source_type").notNull().default("own_site"),
+    title: text("title"),
+    h1: text("h1"),
+    headings: jsonb("headings").$type().notNull().default([]),
+    wordCount: integer("word_count"),
+    fullText: text("full_text"),
+    textHash: text("text_hash"),
+    internalLinks: jsonb("internal_links").$type().notNull().default([]),
+    externalLinks: jsonb("external_links").$type().notNull().default([]),
+    language: text("language"),
+    ticketNumber: text("ticket_number"),
+    metadata: jsonb("metadata").$type().notNull().default({}),
+    extractedAt: timestamp("extracted_at", { withTimezone: true }).notNull().defaultNow(),
+    createdAt: timestamp("created_at", { withTimezone: true }).notNull().defaultNow(),
+    updatedAt: timestamp("updated_at", { withTimezone: true }).notNull().defaultNow()
+  },
+  (table) => [
+    uniqueIndex("content_corpus_items_project_canonical_uidx").on(
+      table.projectKey,
+      table.canonicalUrl
+    ),
+    index("content_corpus_items_domain_idx").on(table.domain),
+    index("content_corpus_items_project_idx").on(table.projectKey),
+    index("content_corpus_items_type_idx").on(table.sourceType),
+    index("content_corpus_items_text_hash_idx").on(table.textHash),
+    index("content_corpus_items_extracted_idx").on(table.extractedAt.desc())
+  ]
+);
+pgTable(
+  "content_quality_runs",
+  {
+    id: uuid("id").primaryKey().defaultRandom(),
+    corpusItemId: uuid("corpus_item_id").references(
+      () => contentCorpusItems.id,
+      { onDelete: "set null" }
+    ),
+    contentSourceId: uuid("content_source_id").references(
+      () => contentSources.id,
+      { onDelete: "set null" }
+    ),
+    projectKey: text("project_key"),
+    siteKey: text("site_key"),
+    url: text("url"),
+    canonicalUrl: text("canonical_url"),
+    tool: text("tool").notNull(),
+    toolVersion: text("tool_version"),
+    score: integer("score"),
+    flags: jsonb("flags").$type().notNull().default([]),
+    report: jsonb("report").$type().notNull().default({}),
+    reportUrl: text("report_url"),
+    ticketNumber: text("ticket_number"),
+    runAt: timestamp("run_at", { withTimezone: true }).notNull().defaultNow(),
+    createdAt: timestamp("created_at", { withTimezone: true }).notNull().defaultNow()
+  },
+  (table) => [
+    index("content_quality_runs_corpus_idx").on(table.corpusItemId),
+    index("content_quality_runs_source_idx").on(table.contentSourceId),
+    index("content_quality_runs_project_run_idx").on(
+      table.projectKey,
+      table.runAt.desc()
+    ),
+    index("content_quality_runs_url_idx").on(table.url)
+  ]
+);
+pgTable(
+  "content_snapshots",
+  {
+    id: uuid("id").primaryKey().defaultRandom(),
+    corpusItemId: uuid("corpus_item_id").references(
+      () => contentCorpusItems.id,
+      { onDelete: "cascade" }
+    ),
+    contentSourceId: uuid("content_source_id").references(
+      () => contentSources.id,
+      { onDelete: "cascade" }
+    ),
+    kind: contentSnapshotKind("kind").notNull(),
+    url: text("url"),
+    canonicalUrl: text("canonical_url"),
+    domain: text("domain"),
+    title: text("title"),
+    fullText: text("full_text"),
+    textHash: text("text_hash"),
+    wordCount: integer("word_count"),
+    headings: jsonb("headings").$type().notNull().default([]),
+    internalLinks: jsonb("internal_links").$type().notNull().default([]),
+    metadata: jsonb("metadata").$type().notNull().default({}),
+    /**
+     * Run that captured this snapshot (managed site content sync, 2026-DASMG-042).
+     * Soft pointer (no Drizzle .references() to avoid a schema import cycle); the
+     * DB foreign key is declared in the migration.
+     */
+    ingestionRunId: uuid("ingestion_run_id"),
+    /** Source git commit / deploy id this snapshot was captured for, when known. */
+    sourceCommit: text("source_commit"),
+    capturedAt: timestamp("captured_at", { withTimezone: true }).notNull().defaultNow(),
+    createdAt: timestamp("created_at", { withTimezone: true }).notNull().defaultNow()
+  },
+  (table) => [
+    index("content_snapshots_corpus_idx").on(
+      table.corpusItemId,
+      table.capturedAt.desc()
+    ),
+    index("content_snapshots_source_idx").on(
+      table.contentSourceId,
+      table.capturedAt.desc()
+    ),
+    index("content_snapshots_text_hash_idx").on(table.textHash),
+    index("content_snapshots_ingestion_run_idx").on(table.ingestionRunId),
+    uniqueIndex("content_snapshots_corpus_hash_uidx").on(
+      table.corpusItemId,
+      table.textHash
+    ),
+    uniqueIndex("content_snapshots_source_hash_uidx").on(
+      table.contentSourceId,
+      table.textHash
+    )
+  ]
+);
 var managedServerOs = pgEnum("managed_server_os", [
   "linux",
   "windows",
@@ -5745,6 +5970,103 @@ pgTable(
     index("github_webhook_secret_owner_idx").on(table.owner)
   ]
 );
+var managedSiteSyncMode = pgEnum("managed_site_sync_mode", [
+  "sitemap",
+  "git",
+  "hybrid",
+  "manual"
+]);
+var contentIngestionTrigger = pgEnum("content_ingestion_trigger", [
+  "manual",
+  "git_push",
+  "deploy",
+  "cron",
+  "sitemap"
+]);
+var contentIngestionStatus = pgEnum("content_ingestion_status", [
+  "queued",
+  "running",
+  "completed",
+  "partial",
+  "failed"
+]);
+var managedSites = pgTable(
+  "managed_sites",
+  {
+    id: uuid("id").primaryKey().defaultRandom(),
+    projectKey: text("project_key").notNull(),
+    siteKey: text("site_key"),
+    name: text("name"),
+    domain: text("domain").notNull(),
+    /** GitHub repo (owner/repo) that publishes this site, when git-driven. */
+    repo: text("repo"),
+    /** Branch whose pushes trigger a sync; null = any branch. */
+    branch: text("branch"),
+    sitemapUrl: text("sitemap_url"),
+    sourceType: contentSourceType("source_type").notNull().default("own_site"),
+    enabled: boolean("enabled").notNull().default(true),
+    syncMode: managedSiteSyncMode("sync_mode").notNull().default("sitemap"),
+    /** Git path -> URL mapping rules used to turn changed files into URLs. */
+    pathUrlMap: jsonb("path_url_map").$type().notNull().default([]),
+    /** Per-run URL cap; null = use the engine default. */
+    maxUrlsPerRun: integer("max_urls_per_run"),
+    /** When true, changed pages are flagged for an MG SEO CLI follow-up (soft handoff). */
+    exportToSeoCli: boolean("export_to_seo_cli").notNull().default(false),
+    lastSyncAt: timestamp("last_sync_at", { withTimezone: true }),
+    /** Soft pointer to the most recent run (no FK to avoid a cycle). */
+    lastRunId: uuid("last_run_id"),
+    metadata: jsonb("metadata").$type().notNull().default({}),
+    createdAt: timestamp("created_at", { withTimezone: true }).notNull().defaultNow(),
+    updatedAt: timestamp("updated_at", { withTimezone: true }).notNull().defaultNow()
+  },
+  (table) => [
+    uniqueIndex("managed_sites_project_domain_uidx").on(
+      table.projectKey,
+      table.domain
+    ),
+    index("managed_sites_repo_idx").on(table.repo),
+    index("managed_sites_enabled_idx").on(table.enabled),
+    index("managed_sites_project_idx").on(table.projectKey)
+  ]
+);
+pgTable(
+  "content_ingestion_runs",
+  {
+    id: uuid("id").primaryKey().defaultRandom(),
+    siteId: uuid("site_id").notNull().references(() => managedSites.id, { onDelete: "cascade" }),
+    projectKey: text("project_key"),
+    triggerType: contentIngestionTrigger("trigger_type").notNull(),
+    status: contentIngestionStatus("status").notNull().default("queued"),
+    /** Trigger.dev run id handling this ingestion, when dispatched. */
+    triggerRunId: text("trigger_run_id"),
+    sourceCommit: text("source_commit"),
+    sourceRef: text("source_ref"),
+    changedUrls: jsonb("changed_urls").$type().notNull().default([]),
+    skippedUrls: jsonb("skipped_urls").$type().notNull().default([]),
+    failedUrls: jsonb("failed_urls").$type().notNull().default([]),
+    limitations: jsonb("limitations").$type().notNull().default([]),
+    /** Corpus item ids touched by this run (for QC / MG SEO CLI handoff). */
+    corpusItemIds: jsonb("corpus_item_ids").$type().notNull().default([]),
+    changedCount: integer("changed_count").notNull().default(0),
+    skippedCount: integer("skipped_count").notNull().default(0),
+    failedCount: integer("failed_count").notNull().default(0),
+    /** Set when changed pages were flagged for an MG SEO CLI follow-up. */
+    qcMarked: boolean("qc_marked").notNull().default(false),
+    error: text("error"),
+    startedAt: timestamp("started_at", { withTimezone: true }),
+    finishedAt: timestamp("finished_at", { withTimezone: true }),
+    createdAt: timestamp("created_at", { withTimezone: true }).notNull().defaultNow()
+  },
+  (table) => [
+    index("content_ingestion_runs_site_idx").on(
+      table.siteId,
+      table.createdAt.desc()
+    ),
+    index("content_ingestion_runs_status_idx").on(table.status),
+    index("content_ingestion_runs_trigger_run_idx").on(table.triggerRunId),
+    index("content_ingestion_runs_project_idx").on(table.projectKey)
+  ]
+);
 var GITHUB_API = "https://api.github.com";
 var ReleaseProfileQuickCreateError = class extends Error {
   constructor(message, code = "INTERNAL_SERVER_ERROR") {
@@ -7075,7 +7397,7 @@ function getEncryptionKey() {
     throw new Error("ENCRYPTION_KEY must be a 64-character hex string");
   return buf;
 }
-function encrypt(text7) {
+function encrypt(text9) {
   const key = getEncryptionKey();
   const iv = randomBytes(ENC_IV_LENGTH);
   const cipher = createCipheriv(
@@ -7083,7 +7405,7 @@ function encrypt(text7) {
     new Uint8Array(key),
     new Uint8Array(iv)
   );
-  let encrypted = cipher.update(text7, "utf8", "hex");
+  let encrypted = cipher.update(text9, "utf8", "hex");
   encrypted += cipher.final("hex");
   const authTag = cipher.getAuthTag();
   return Buffer.concat([
@@ -7803,10 +8125,10 @@ async function r2GetObjectRange(bucket, key, range) {
     const body = result.Body;
     if (!body?.transformToString)
       throw new Error("R2 returned no readable body");
-    const text7 = await body.transformToString();
+    const text9 = await body.transformToString();
     const header = `# range: bytes ${range.offset}-${end} of ${size} (${effectiveLen} bytes)`;
     return `${header}
-${text7}`;
+${text9}`;
   } catch (e) {
     throw r2WrapError(bucket, key, e);
   }
@@ -8169,15 +8491,15 @@ async function sftpRead(opts, filePath, proxy, options) {
             clearTimeout(timer);
             cleanup?.();
             cleanup = void 0;
-            const text7 = Buffer.concat(
+            const text9 = Buffer.concat(
               chunks.map((ch) => new Uint8Array(ch))
             ).toString("utf-8");
             if (!isWholeFileRequest) {
               const header = `# range: bytes ${offset}-${offset + effectiveLen - 1} of ${total} (${effectiveLen} bytes)`;
               resolve(`${header}
-${text7}`);
+${text9}`);
             } else {
-              resolve(text7);
+              resolve(text9);
             }
           });
           rs.on("error", (e) => {
@@ -8211,7 +8533,27 @@ var NO_FOOTER_TOOLS = /* @__PURE__ */ new Set();
 var RAW_JSON_TOOLS = /* @__PURE__ */ new Set([
   "get_mg_dashboard_commits",
   "extract_article",
-  "research_topic"
+  "research_topic",
+  // Content registry / corpus (2026-DASMG-041) — compact JSON consumed verbatim.
+  "save_content_source",
+  "get_content_source_by_url",
+  "get_content_source_by_id",
+  "search_content_sources",
+  "save_research_pack",
+  "get_research_pack",
+  "save_content_corpus_item",
+  "list_project_content_corpus",
+  "get_content_context_for_project",
+  "save_content_quality_run",
+  "list_content_snapshots",
+  "link_content_pack_to_ticket",
+  "prune_content_snapshots",
+  // Managed site content sync (2026-DASMG-042) — compact JSON consumed verbatim.
+  "register_managed_site",
+  "update_managed_site",
+  "list_managed_sites",
+  "trigger_content_sync",
+  "get_content_sync_status"
 ]);
 var TOOL_CACHE_TTL_MS = {
   "list-servers": 6e4,
@@ -8312,11 +8654,11 @@ function getKnownContainers(serverId, maxAgeMs = 5 * 6e4) {
   if (Date.now() - e.capturedAt > maxAgeMs) return void 0;
   return e.names;
 }
-function truncateForLLM(text7, maxBytes) {
-  const totalBytes = Buffer.byteLength(text7, "utf8");
+function truncateForLLM(text9, maxBytes) {
+  const totalBytes = Buffer.byteLength(text9, "utf8");
   if (totalBytes <= maxBytes)
-    return { text: text7, truncated: false, totalBytes, shownBytes: totalBytes };
-  const buf = Buffer.from(text7, "utf8");
+    return { text: text9, truncated: false, totalBytes, shownBytes: totalBytes };
+  const buf = Buffer.from(text9, "utf8");
   let cut = maxBytes;
   while (cut > 0 && (buf[cut] & 192) === 128) cut--;
   const head = buf.subarray(0, cut).toString("utf8");
@@ -8347,10 +8689,10 @@ function postprocessResult(result, meta) {
   if (!result.content?.length) return result;
   if (RAW_JSON_TOOLS.has(meta.toolName)) return result;
   const block = result.content[0];
-  let text7 = String(block.text ?? "");
-  const trunc = truncateForLLM(text7, RESPONSE_MAX_BYTES);
+  let text9 = String(block.text ?? "");
+  const trunc = truncateForLLM(text9, RESPONSE_MAX_BYTES);
   if (trunc.truncated) {
-    text7 = trunc.text + "\n\n... " + buildTruncationHint(
+    text9 = trunc.text + "\n\n... " + buildTruncationHint(
       meta.toolName,
       meta.args,
       trunc.totalBytes,
@@ -8364,11 +8706,11 @@ function postprocessResult(result, meta) {
     const parts = [`took ${tookStr}`, sizeStr];
     if (meta.serverIdLabel) parts.push(`server: ${meta.serverIdLabel}`);
     if (meta.cached) parts.push("cached");
-    text7 = `${text7}
+    text9 = `${text9}
 
 [${parts.join(", ")}]`;
   }
-  return { ...result, content: [{ ...block, text: text7 }] };
+  return { ...result, content: [{ ...block, text: text9 }] };
 }
 function buildPipelineScript(commands, shell, marker, stopOnError) {
   if (shell === "powershell") {
@@ -10420,6 +10762,462 @@ var TOOLS = [
       }
     }
   },
+  // ----- Content registry / corpus persistence (2026-DASMG-041) -----
+  {
+    name: "save_content_source",
+    description: "Persist (dedupe + upsert) ONE fetched source \u2014 the durable companion to extract_article (2026-DASMG-041). Pass a `url` (re-extracts internally) and/or already-extracted fields (title/summary/facts/fullText). Dedupes on canonical URL then text hash, and snapshots the previous version into content_snapshots when the text changed. External sources keep their `source_type` so they stay labelled as evidence, never own content. Returns the stored id + a compact summary + whether it was created/updated/snapshotted.",
+    inputSchema: {
+      type: "object",
+      properties: {
+        url: { type: "string", description: "Source URL (required)." },
+        canonicalUrl: { type: "string", description: "Canonical URL used for dedupe." },
+        sourceType: {
+          type: "string",
+          enum: [
+            "own_site",
+            "client_site",
+            "external_source",
+            "competitor",
+            "news",
+            "government",
+            "supplier",
+            "unknown"
+          ],
+          description: "Provenance label. own_site/client_site = managed content; the rest are evidence only."
+        },
+        title: { type: "string", description: "Title." },
+        author: { type: "string", description: "Author." },
+        publishedAt: { type: "string", description: "Publish date (any parseable format)." },
+        summary: { type: "string", description: "Short factual summary." },
+        facts: { type: "array", items: { type: "string" }, description: "Key facts." },
+        limitations: { type: "array", items: { type: "string" } },
+        fullText: { type: "string", description: "Extracted body text." },
+        language: { type: "string", description: "Language code." },
+        projectKey: { type: "string", description: "Project/site identifier." },
+        siteKey: { type: "string", description: "Site identifier." },
+        ticketNumber: { type: "string", description: "Related ticket number." },
+        metadata: { type: "object" },
+        reExtract: {
+          type: "boolean",
+          description: "Force a re-fetch even when fields are supplied."
+        }
+      },
+      required: ["url"]
+    }
+  },
+  {
+    name: "get_content_source_by_url",
+    description: "Look up a stored source by URL or canonical URL (2026-DASMG-041). Compact by default (id + summary + counts); set `includeFullText` for the body. Use to answer 'was this link already scanned?' before re-extracting.",
+    inputSchema: {
+      type: "object",
+      properties: {
+        url: { type: "string", description: "URL or canonical URL." },
+        includeFullText: { type: "boolean", description: "Return full_text." }
+      },
+      required: ["url"]
+    }
+  },
+  {
+    name: "get_content_source_by_id",
+    description: "Fetch a stored source by id (2026-DASMG-041). Compact by default; set `includeFullText` for the body.",
+    inputSchema: {
+      type: "object",
+      properties: {
+        id: { type: "string", description: "content_sources.id (uuid)." },
+        includeFullText: { type: "boolean", description: "Return full_text." }
+      },
+      required: ["id"]
+    }
+  },
+  {
+    name: "search_content_sources",
+    description: "Full-text + filter search over stored sources (2026-DASMG-041): query (title/summary/body), domain, sourceType, projectKey, ticketNumber, dateFrom/dateTo. Compact, paginated results. Use to reuse evidence already in the registry.",
+    inputSchema: {
+      type: "object",
+      properties: {
+        query: { type: "string", description: "Free-text search (title/summary/body)." },
+        domain: { type: "string", description: "Exact domain filter." },
+        sourceType: {
+          type: "string",
+          enum: [
+            "own_site",
+            "client_site",
+            "external_source",
+            "competitor",
+            "news",
+            "government",
+            "supplier",
+            "unknown"
+          ]
+        },
+        projectKey: { type: "string", description: "Project filter." },
+        ticketNumber: { type: "string", description: "Ticket filter." },
+        dateFrom: { type: "string", description: "Last-seen lower bound (ISO/date)." },
+        dateTo: { type: "string", description: "Last-seen upper bound (ISO/date)." },
+        limit: { type: "integer", description: "Max rows (default 20, max 100)." },
+        offset: { type: "integer", description: "Pagination offset." },
+        includeFullText: { type: "boolean", description: "Return full_text per row." }
+      }
+    }
+  },
+  {
+    name: "save_research_pack",
+    description: "Persist a research_topic evidence pack (2026-DASMG-041): cross-checked verified facts, conflicts, unique claims, open questions. Pass an already-built `pack`, or `urls`/`topic` to research then save. Each source is deduped into content_sources and linked via the join table, so packs can be reused without re-crawling. Returns the pack id + compact counts + linked source ids.",
+    inputSchema: {
+      type: "object",
+      properties: {
+        pack: { type: "object", description: "A research_topic result payload." },
+        urls: {
+          type: "array",
+          items: { type: "string" },
+          description: "URLs to research if no pack is given."
+        },
+        topic: { type: "string", description: "Topic to research if no pack is given." },
+        locale: { type: "string", description: "Locale of the pack." },
+        projectKey: { type: "string", description: "Project identifier." },
+        siteKey: { type: "string", description: "Site identifier." },
+        ticketNumber: { type: "string", description: "Related ticket number." },
+        sourceType: {
+          type: "string",
+          enum: [
+            "own_site",
+            "client_site",
+            "external_source",
+            "competitor",
+            "news",
+            "government",
+            "supplier",
+            "unknown"
+          ]
+        },
+        metadata: { type: "object" }
+      }
+    }
+  },
+  {
+    name: "get_research_pack",
+    description: "Fetch a stored evidence pack by id (2026-DASMG-041). Compact by default (counts + provenance); set `includeSources` / `includeClaims` for the full arrays.",
+    inputSchema: {
+      type: "object",
+      properties: {
+        id: { type: "string", description: "content_evidence_packs.id (uuid)." },
+        includeSources: { type: "boolean", description: "Return the source list." },
+        includeClaims: {
+          type: "boolean",
+          description: "Return verified facts / conflicts / unique claims."
+        }
+      },
+      required: ["id"]
+    }
+  },
+  {
+    name: "save_content_corpus_item",
+    description: "Upsert ONE own/managed page into the corpus (2026-DASMG-041), keyed by (projectKey, canonical url) \u2014 the current state of content you own. Pass a `url` (re-extracts) and/or title/h1/headings/fullText/internalLinks. Snapshots the previous version when the text changed. Use this for your/your client's pages, not external evidence.",
+    inputSchema: {
+      type: "object",
+      properties: {
+        projectKey: { type: "string", description: "Project identifier (required)." },
+        url: { type: "string", description: "Page URL (required)." },
+        siteKey: { type: "string", description: "Site identifier." },
+        canonicalUrl: { type: "string", description: "Canonical URL." },
+        sourceType: {
+          type: "string",
+          enum: [
+            "own_site",
+            "client_site",
+            "external_source",
+            "competitor",
+            "news",
+            "government",
+            "supplier",
+            "unknown"
+          ],
+          description: "Defaults to own_site."
+        },
+        pagePath: { type: "string", description: "Path within the site." },
+        title: { type: "string", description: "Title." },
+        h1: { type: "string", description: "Main heading." },
+        headings: {
+          type: "array",
+          items: { type: "object" },
+          description: "Headings ({level,text})."
+        },
+        wordCount: { type: "integer", description: "Word count." },
+        fullText: { type: "string", description: "Full page text." },
+        internalLinks: {
+          type: "array",
+          items: { type: "object" },
+          description: "Internal links ({url,text})."
+        },
+        externalLinks: { type: "array", items: { type: "object" } },
+        language: { type: "string", description: "Language code." },
+        ticketNumber: { type: "string", description: "Related ticket number." },
+        metadata: { type: "object" },
+        reExtract: {
+          type: "boolean",
+          description: "Force a re-fetch even when fields are supplied."
+        }
+      },
+      required: ["projectKey", "url"]
+    }
+  },
+  {
+    name: "list_project_content_corpus",
+    description: "List corpus items for a project (2026-DASMG-041), filter by domain/sourceType/query. Compact, paginated. Use to see content coverage and reuse own pages.",
+    inputSchema: {
+      type: "object",
+      properties: {
+        projectKey: { type: "string", description: "Project identifier (required)." },
+        domain: { type: "string", description: "Domain filter." },
+        sourceType: {
+          type: "string",
+          enum: [
+            "own_site",
+            "client_site",
+            "external_source",
+            "competitor",
+            "news",
+            "government",
+            "supplier",
+            "unknown"
+          ]
+        },
+        query: { type: "string", description: "Free-text search." },
+        limit: { type: "integer", description: "Max rows (default 20)." },
+        offset: { type: "integer", description: "Pagination offset." },
+        includeFullText: { type: "boolean", description: "Return full_text per row." }
+      },
+      required: ["projectKey"]
+    }
+  },
+  {
+    name: "get_content_context_for_project",
+    description: "Aggregate context for a content agent (2026-DASMG-041): recent corpus items, recent evidence packs and related sources for a project (optionally focused by topic or url), plus totals. One call to brief an agent on what already exists before it writes.",
+    inputSchema: {
+      type: "object",
+      properties: {
+        projectKey: { type: "string", description: "Project identifier (required)." },
+        topic: { type: "string", description: "Optional topic focus." },
+        url: { type: "string", description: "Optional url focus." },
+        limit: { type: "integer", description: "Max items per section (default 10)." }
+      },
+      required: ["projectKey"]
+    }
+  },
+  {
+    name: "save_content_quality_run",
+    description: "Store a reference to an MG SEO CLI / audit run (2026-DASMG-041): score, flags and report, attached to a corpus item, source or url. MG Dashboard stores the run; the CLI does the scoring (the CLI stays the QC layer).",
+    inputSchema: {
+      type: "object",
+      properties: {
+        tool: { type: "string", description: "Tool name, e.g. mg-seo (required)." },
+        corpusItemId: { type: "string", description: "content_corpus_items.id." },
+        contentSourceId: { type: "string", description: "content_sources.id." },
+        projectKey: { type: "string", description: "Project identifier." },
+        siteKey: { type: "string", description: "Site identifier." },
+        url: { type: "string", description: "Audited URL." },
+        canonicalUrl: { type: "string", description: "Canonical URL." },
+        toolVersion: { type: "string", description: "Tool version." },
+        score: { type: "integer", description: "Overall score 0-100." },
+        flags: {
+          type: "array",
+          items: { type: "object" },
+          description: "Flags ({code,severity,message})."
+        },
+        report: { type: "object", description: "Full report payload." },
+        reportUrl: { type: "string", description: "Link to the full report." },
+        ticketNumber: { type: "string", description: "Related ticket number." },
+        runAt: { type: "string", description: "Run timestamp (ISO)." }
+      },
+      required: ["tool"]
+    }
+  },
+  {
+    name: "list_content_snapshots",
+    description: "List historical snapshots for a corpus item or source (2026-DASMG-041) by id, url or canonical url, newest first. Compact; set `includeFullText` for the captured body. Use to see how content changed over time.",
+    inputSchema: {
+      type: "object",
+      properties: {
+        corpusItemId: { type: "string", description: "content_corpus_items.id." },
+        contentSourceId: { type: "string", description: "content_sources.id." },
+        url: { type: "string", description: "URL to match." },
+        canonicalUrl: { type: "string", description: "Canonical URL to match." },
+        limit: { type: "integer", description: "Max rows (default 20)." },
+        includeFullText: { type: "boolean", description: "Return captured full_text." }
+      }
+    }
+  },
+  {
+    name: "link_content_pack_to_ticket",
+    description: "Attach a stored evidence pack to a ticket number (2026-DASMG-041).",
+    inputSchema: {
+      type: "object",
+      properties: {
+        packId: { type: "string", description: "content_evidence_packs.id (uuid)." },
+        ticketNumber: { type: "string", description: "Ticket number." }
+      },
+      required: ["packId", "ticketNumber"]
+    }
+  },
+  {
+    name: "prune_content_snapshots",
+    description: "Retention cleanup (2026-DASMG-041): keep the newest N snapshots per item and drop the rest older than the retention window. Defaults: keepPerItem 20, olderThanDays 365. Returns how many were pruned.",
+    inputSchema: {
+      type: "object",
+      properties: {
+        keepPerItem: { type: "integer", description: "Snapshots to keep per item (default 20)." },
+        olderThanDays: {
+          type: "integer",
+          description: "Only prune snapshots older than this many days (default 365)."
+        }
+      }
+    }
+  },
+  {
+    name: "register_managed_site",
+    description: "Register (or upsert by projectKey+domain) a site for content sync (2026-DASMG-042): domain, optional repo/branch + sitemapUrl, syncMode and a git pathUrlMap. Optional + additive \u2014 external/unmanaged sites still work via extract_article without this. Enables git-push/cron/manual ingestion that re-indexes only CHANGED pages into the corpus (hash-skips unchanged ones).",
+    inputSchema: {
+      type: "object",
+      properties: {
+        projectKey: { type: "string", description: "Project identifier (required)." },
+        domain: { type: "string", description: "Site domain, e.g. example.nl (required)." },
+        siteKey: { type: "string", description: "Optional site identifier within the project." },
+        name: { type: "string", description: "Human-friendly site name." },
+        repo: { type: "string", description: "GitHub repo owner/repo that publishes the site (for git pushes)." },
+        branch: { type: "string", description: "Branch whose pushes trigger a sync; omit for any branch." },
+        sitemapUrl: { type: "string", description: "Sitemap URL for sitemap/hybrid detection." },
+        sourceType: {
+          type: "string",
+          enum: [
+            "own_site",
+            "client_site",
+            "external_source",
+            "competitor",
+            "news",
+            "government",
+            "supplier",
+            "unknown"
+          ],
+          description: "Provenance label for produced corpus items (default own_site)."
+        },
+        enabled: { type: "boolean", description: "Whether syncs run (default true)." },
+        syncMode: {
+          type: "string",
+          enum: ["sitemap", "git", "hybrid", "manual"],
+          description: "Change detection: sitemap (lastmod/hash), git (changed paths via pathUrlMap), hybrid (both), manual (only on trigger_content_sync)."
+        },
+        pathUrlMap: {
+          type: "array",
+          items: { type: "object" },
+          description: "Git path -> URL rules ([{pathPrefix,urlPrefix}])."
+        },
+        maxUrlsPerRun: { type: "integer", description: "Per-run URL cap (default 200)." },
+        exportToSeoCli: {
+          type: "boolean",
+          description: "Flag changed pages for an MG SEO CLI follow-up (soft handoff)."
+        },
+        metadata: { type: "object" }
+      },
+      required: ["projectKey", "domain"]
+    }
+  },
+  {
+    name: "update_managed_site",
+    description: "Update an existing managed site by id (2026-DASMG-042). Only the provided fields change \u2014 enable/disable, switch syncMode, set sitemapUrl/pathUrlMap, etc.",
+    inputSchema: {
+      type: "object",
+      properties: {
+        id: { type: "string", description: "managed_sites.id (uuid, required)." },
+        name: { type: "string", description: "Human-friendly site name." },
+        domain: { type: "string", description: "Site domain." },
+        siteKey: { type: "string", description: "Site identifier within the project." },
+        repo: { type: "string", description: "GitHub repo owner/repo." },
+        branch: { type: "string", description: "Branch whose pushes trigger a sync." },
+        sitemapUrl: { type: "string", description: "Sitemap URL." },
+        sourceType: {
+          type: "string",
+          enum: [
+            "own_site",
+            "client_site",
+            "external_source",
+            "competitor",
+            "news",
+            "government",
+            "supplier",
+            "unknown"
+          ]
+        },
+        enabled: { type: "boolean", description: "Enable/disable syncs." },
+        syncMode: {
+          type: "string",
+          enum: ["sitemap", "git", "hybrid", "manual"]
+        },
+        pathUrlMap: { type: "array", items: { type: "object" } },
+        maxUrlsPerRun: { type: "integer", description: "Per-run URL cap." },
+        exportToSeoCli: { type: "boolean" },
+        metadata: { type: "object" }
+      },
+      required: ["id"]
+    }
+  },
+  {
+    name: "list_managed_sites",
+    description: "List managed sites (2026-DASMG-042), filter by projectKey/domain/repo/enabledOnly. Compact, paginated.",
+    inputSchema: {
+      type: "object",
+      properties: {
+        projectKey: { type: "string", description: "Project filter." },
+        domain: { type: "string", description: "Domain filter." },
+        repo: { type: "string", description: "Repo filter (owner/repo)." },
+        enabledOnly: { type: "boolean", description: "Only enabled sites." },
+        limit: { type: "integer", description: "Max rows (default 50, max 200)." },
+        offset: { type: "integer", description: "Pagination offset." }
+      }
+    }
+  },
+  {
+    name: "trigger_content_sync",
+    description: "Start a content ingestion run for a managed site (2026-DASMG-042) by siteId, or projectKey+domain. Detects changed pages (sitemap/git/hash), re-indexes only changes into the corpus and snapshots the previous version. Returns the runId; poll get_content_sync_status for results.",
+    inputSchema: {
+      type: "object",
+      properties: {
+        siteId: { type: "string", description: "managed_sites.id (uuid)." },
+        projectKey: { type: "string", description: "Project identifier (with domain)." },
+        domain: { type: "string", description: "Site domain (with projectKey)." },
+        triggerType: {
+          type: "string",
+          enum: ["manual", "git_push", "deploy", "cron", "sitemap"],
+          description: "Trigger label recorded on the run (default manual)."
+        },
+        sourceCommit: { type: "string", description: "Source git commit/deploy id, when known." },
+        sourceRef: { type: "string", description: "Source git ref, when known." },
+        urls: {
+          type: "array",
+          items: { type: "string" },
+          description: "Restrict the run to these URLs (skips detection)."
+        },
+        changedPaths: {
+          type: "array",
+          items: { type: "string" },
+          description: "Changed repo paths to map to URLs via pathUrlMap."
+        }
+      }
+    }
+  },
+  {
+    name: "get_content_sync_status",
+    description: "Read ingestion run status (2026-DASMG-042): by runId (that run), or by siteId / projectKey+domain (recent runs). Returns changed/skipped/failed counts, the changed URLs, limitations and the site summary.",
+    inputSchema: {
+      type: "object",
+      properties: {
+        runId: { type: "string", description: "content_ingestion_runs.id (uuid)." },
+        siteId: { type: "string", description: "managed_sites.id (uuid)." },
+        projectKey: { type: "string", description: "Project identifier (with domain)." },
+        domain: { type: "string", description: "Site domain (with projectKey)." },
+        limit: { type: "integer", description: "Max recent runs to return (default 10)." }
+      }
+    }
+  },
   {
     name: "search-team-memory",
     description: "Search the team's ENTIRE past Cursor history (every developer, every project) for how something was handled before. Use this FIRST, before investigating from scratch, whenever you: hit a non-trivial bug or error, are about to build something that may have been done before, need a project-specific convention/gotcha, or the user asks 'have we done X', 'how did we fix Y', or 'did we solve this already'. Hybrid semantic + keyword search over mirrored conversations. Returns ranked past chats with repo, date, a solution snippet and a similarity score. Phrase the query as the problem in natural language (e.g. 'release pipeline PM2 deploy fails with module not found').",
@@ -10520,7 +11318,7 @@ var TOOLS = [
   // ----- Trigger.dev -----
   ...TRIGGER_TOOLS
 ];
-var MCP_VERSION = "7.0.5";
+var MCP_VERSION = "7.3.0";
 async function handleListTools() {
   if (!authContext) return { tools: TOOLS };
   const allowedTools = authContext.allowedTools;
@@ -10723,6 +11521,80 @@ async function executeToolCall(name, a, _serverId) {
           content: [{ type: "text", text: JSON.stringify(data, null, 2) }]
         };
       }
+      // ----- Content registry / corpus persistence (2026-DASMG-041) -----
+      // All content-registry tools proxy to one dispatcher route; the tool name
+      // is the dispatcher `action` and the args are forwarded verbatim as
+      // `params` (the route validates them with the matching Zod schema).
+      case "save_content_source":
+      case "get_content_source_by_url":
+      case "get_content_source_by_id":
+      case "search_content_sources":
+      case "save_research_pack":
+      case "get_research_pack":
+      case "save_content_corpus_item":
+      case "list_project_content_corpus":
+      case "get_content_context_for_project":
+      case "save_content_quality_run":
+      case "list_content_snapshots":
+      case "link_content_pack_to_ticket":
+      case "prune_content_snapshots": {
+        const res = await fetch(`${dashboardBaseUrl}/api/tools/content-registry`, {
+          method: "POST",
+          headers: {
+            "content-type": "application/json",
+            authorization: `Bearer ${apiKey}`
+          },
+          body: JSON.stringify({ action: name, params: a })
+        });
+        if (!res.ok) {
+          const detail = await res.text().catch(() => "");
+          return {
+            content: [
+              {
+                type: "text",
+                text: `Error: ${name} failed (${res.status}). ${detail.slice(0, 300)}`
+              }
+            ]
+          };
+        }
+        const data = await res.json();
+        return {
+          content: [{ type: "text", text: JSON.stringify(data, null, 2) }]
+        };
+      }
+      // ----- Managed site content sync (2026-DASMG-042) -----
+      // The 5 PUBLIC managed-site tools proxy to one dispatcher route; the tool
+      // name is the dispatcher `action`, args forwarded verbatim as `params`
+      // (the route validates them and only allows the public actions).
+      case "register_managed_site":
+      case "update_managed_site":
+      case "list_managed_sites":
+      case "trigger_content_sync":
+      case "get_content_sync_status": {
+        const res = await fetch(`${dashboardBaseUrl}/api/tools/managed-site`, {
+          method: "POST",
+          headers: {
+            "content-type": "application/json",
+            authorization: `Bearer ${apiKey}`
+          },
+          body: JSON.stringify({ action: name, params: a })
+        });
+        if (!res.ok) {
+          const detail = await res.text().catch(() => "");
+          return {
+            content: [
+              {
+                type: "text",
+                text: `Error: ${name} failed (${res.status}). ${detail.slice(0, 300)}`
+              }
+            ]
+          };
+        }
+        const data = await res.json();
+        return {
+          content: [{ type: "text", text: JSON.stringify(data, null, 2) }]
+        };
+      }
       // ----- Team memory -----
       case "search-team-memory": {
         const query = typeof a.query === "string" ? a.query.trim() : "";
@@ -10764,7 +11636,7 @@ async function executeToolCall(name, a, _serverId) {
             ]
           };
         }
-        const lines = data.hits.map((hit, index5) => {
+        const lines = data.hits.map((hit, index7) => {
           const when = hit.lastMessageAt ? new Date(hit.lastMessageAt).toISOString().slice(0, 10) : "unknown date";
           const repo = hit.repo ?? "unknown repo";
           const sim = hit.similarity !== null ? ` \xB7 ${(hit.similarity * 100).toFixed(0)}% match` : "";
@@ -10774,7 +11646,7 @@ async function executeToolCall(name, a, _serverId) {
             ...Array.isArray(hit.tech) ? hit.tech.slice(0, 4) : []
           ].filter(Boolean);
           const tags = facets.length > 0 ? ` \xB7 ${facets.join(", ")}` : "";
-          return `${index5 + 1}. [${repo} \xB7 ${when}${sim}${tags}] ${hit.title}
+          return `${index7 + 1}. [${repo} \xB7 ${when}${sim}${tags}] ${hit.title}
    id: ${hit.id}
    ${(hit.summary ?? hit.snippet).replace(/\s+/g, " ").slice(0, 500)}`;
         });
@@ -11959,8 +12831,8 @@ ${sample.join("\n")}${files.length > 5 ? `
             };
           const filtered = sortRows(applyFilter(only.rows));
           if (format === "json") {
-            const text8 = filtered.map((c) => JSON.stringify(c)).join("\n") || "(no containers)";
-            return { content: [{ type: "text", text: text8 }] };
+            const text10 = filtered.map((c) => JSON.stringify(c)).join("\n") || "(no containers)";
+            return { content: [{ type: "text", text: text10 }] };
           }
           if (groupByProject) {
             const groups = /* @__PURE__ */ new Map();
@@ -11987,8 +12859,8 @@ ${sample.join("\n")}${files.length > 5 ? `
           }
           const header = `${"NAMES".padEnd(36)}  ${"IMAGE".padEnd(40)}  ${"STATUS".padEnd(24)}  ${"HEALTH".padEnd(10)}  ${"PORTS".padEnd(40)}  PROJECT`;
           const body = filtered.map((r) => `${fmtRow(r)}  ${r.Project || ""}`);
-          const text7 = filtered.length === 0 ? "(no containers match)" : [header, ...body].join("\n");
-          return { content: [{ type: "text", text: text7 }] };
+          const text9 = filtered.length === 0 ? "(no containers match)" : [header, ...body].join("\n");
+          return { content: [{ type: "text", text: text9 }] };
         }
         if (format === "json") {
           const lines = [];