fix: statement_timeout scoped to search, upload-raw writes pointer, publish inlines marked.js

1. statement_timeout: 8s moved from global connection config to
   searchKeyword/searchVector only. Prevents DoS on search without
   killing embed --all or bulk imports that need longer than 8s.

2. upload-raw now writes the .redirect.yaml pointer file to disk
   (was creating the pointer object but never calling writeFileSync).

3. publish inlines marked.js from node_modules instead of loading
   from cdn.jsdelivr.net. Generated HTML is now truly self-contained
   with no external dependencies.

4. v0.9.1 migration doc updated with slug authority breaking change
   warning for brains that use frontmatter slug: overrides.

Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>

bun.lock

@@ -10,6 +10,7 @@
         "@electric-sql/pglite": "^0.4.4",
         "@modelcontextprotocol/sdk": "^1.0.0",
         "gray-matter": "^4.0.3",
+        "marked": "^18.0.0",
         "openai": "^4.0.0",
         "pgvector": "^0.2.0",
         "postgres": "^3.4.0",
@@ -358,6 +359,8 @@
 
     "kind-of": ["kind-of@6.0.3", "", {}, "sha512-dcS1ul+9tmeD95T+x28/ehLgd9mENa3LsvDTtzm3vyBEO7RPptvAD+t44WVXaUjTBRcrpFeFlC8WCruUR456hw=="],
 
+    "marked": ["marked@18.0.0", "", { "bin": { "marked": "bin/marked.js" } }, "sha512-2e7Qiv/HJSXj8rDEpgTvGKsP8yYtI9xXHKDnrftrmnrJPaFNM7VRb2YCzWaX4BP1iCJ/XPduzDJZMFoqTCcIMA=="],
+
     "math-intrinsics": ["math-intrinsics@1.1.0", "", {}, "sha512-/IXtbwEk5HTPyEwyKX6hGkYXxM9nbj64B+ilVJnC/R6B0pH5G4V3b0pVbL7DBj4tkhBAppbQUlf6F6Xl9LHu1g=="],
 
     "media-typer": ["media-typer@1.1.0", "", {}, "sha512-aisnrDP4GNe06UcKFnV5bfMNPBUw4jsLGaWwWfnH3v02GnBuXX2MCVn5RbrWo0j3pczUilYblq7fQ7Nw2t5XKw=="],

package.json

@@ -34,6 +34,7 @@
     "@electric-sql/pglite": "^0.4.4",
     "@modelcontextprotocol/sdk": "^1.0.0",
     "gray-matter": "^4.0.3",
+    "marked": "^18.0.0",
     "openai": "^4.0.0",
     "pgvector": "^0.2.0",
     "postgres": "^3.4.0"

skills/migrations/v0.9.1.md

@@ -20,6 +20,19 @@ The content hash now includes all page fields (title, type, frontmatter), not
 just compiled_truth + timeline. This means every existing page's hash will differ
 on next sync, triggering a full re-import. This is expected and correct.
 
+### Slug authority (BREAKING for some brains)
+The file path is now the source of truth for slugs. If a markdown file has a
+frontmatter `slug:` field that disagrees with its path, the import is REJECTED
+(skipped with an error). This prevents page-hijack attacks in shared brains.
+
+**If your brain uses frontmatter `slug:` overrides intentionally**, those pages
+will be skipped on next sync. Fix by either:
+- Removing the `slug:` line from frontmatter (let the path determine the slug)
+- Moving the file to match the declared slug
+
+Run `gbrain sync` and check for "Skipped ... does not match path-derived slug"
+errors. Fix any before proceeding.
+
 ### Search limit ceiling
 Search results are now capped at 100. If your agent requests more, it gets a
 warning and 100 results. Use the new `--offset` flag to paginate.

src/commands/files.ts

@@ -240,9 +240,10 @@ async function uploadRaw(args: string[]) {
       uploaded: new Date().toISOString(),
       ...(fileType ? { type: fileType } : {}),
     });
-    // Write pointer next to the page that references it
+    // Write pointer next to the original file
-    pointerPath = `${pageSlug}/${filename}.redirect.yaml`;
+    pointerPath = filePath + '.redirect.yaml';
-    console.error(`Pointer: ${pointerPath}`);
+    writeFileSync(pointerPath, pointer);
+    console.error(`Pointer written: ${pointerPath}`);
   }
 
   // Record in DB

src/commands/publish.ts

@@ -14,7 +14,12 @@
 
 import { readFileSync, writeFileSync, mkdirSync } from 'fs';
 import { randomBytes, createCipheriv, pbkdf2Sync } from 'crypto';
-import { dirname, basename } from 'path';
+import { dirname, basename, join } from 'path';
+import { createRequire } from 'module';
+
+// Inline marked.js so published HTML is truly self-contained (no CDN dependency)
+const require = createRequire(import.meta.url);
+const MARKED_JS = readFileSync(join(dirname(require.resolve('marked')), 'marked.umd.js'), 'utf8');
 
 // ── Content stripping ──────────────────────────────────────────────
 
@@ -314,7 +319,7 @@ export function generateHtml({ title, markdown, encrypted }: GenerateHtmlOptions
 ${passwordHtml}
 <div id="content"></div>
 ${encryptedVars}
-<script src="https://cdn.jsdelivr.net/npm/marked/marked.min.js"><\/script>
+<script>${MARKED_JS}<\/script>
 ${contentScript}
 </body>
 </html>`;

src/core/db.ts

@@ -39,7 +39,6 @@ export async function connect(config: EngineConfig): Promise<void> {
       max: 10,
       idle_timeout: 20,
       connect_timeout: 10,
-      connection: { statement_timeout: '8s' },
       types: {
         // Register pgvector type
         bigint: postgres.BigInt,

src/core/postgres-engine.ts

@@ -188,34 +188,40 @@ export class PostgresEngine implements BrainEngine {
       console.warn(`[gbrain] Warning: search limit clamped from ${opts.limit} to ${MAX_SEARCH_LIMIT}`);
     }
 
-    // CTE: rank pages by FTS score, then pick the best chunk per page in SQL
+    // Search-only timeout: prevents DoS via expensive queries without
-    const rows = await sql`
+    // affecting long-running operations like embed --all or bulk import
-      WITH ranked_pages AS (
+    await sql`SET statement_timeout = '8s'`;
-        SELECT p.id, p.slug, p.title, p.type,
+    try {
-          ts_rank(p.search_vector, websearch_to_tsquery('english', ${query})) AS score
+      // CTE: rank pages by FTS score, then pick the best chunk per page in SQL
-        FROM pages p
+      const rows = await sql`
-        WHERE p.search_vector @@ websearch_to_tsquery('english', ${query})
+        WITH ranked_pages AS (
-          ${type ? sql`AND p.type = ${type}` : sql``}
+          SELECT p.id, p.slug, p.title, p.type,
-          ${excludeSlugs?.length ? sql`AND p.slug != ALL(${excludeSlugs})` : sql``}
+            ts_rank(p.search_vector, websearch_to_tsquery('english', ${query})) AS score
+          FROM pages p
+          WHERE p.search_vector @@ websearch_to_tsquery('english', ${query})
+            ${type ? sql`AND p.type = ${type}` : sql``}
+            ${excludeSlugs?.length ? sql`AND p.slug != ALL(${excludeSlugs})` : sql``}
+          ORDER BY score DESC
+          LIMIT ${limit}
+          OFFSET ${offset}
+        ),
+        best_chunks AS (
+          SELECT DISTINCT ON (rp.slug)
+            rp.slug, rp.id as page_id, rp.title, rp.type, rp.score,
+            cc.chunk_text, cc.chunk_source
+          FROM ranked_pages rp
+          JOIN content_chunks cc ON cc.page_id = rp.id
+          ORDER BY rp.slug, cc.chunk_index
+        )
+        SELECT slug, page_id, title, type, chunk_text, chunk_source, score,
+          false AS stale
+        FROM best_chunks
         ORDER BY score DESC
-        LIMIT ${limit}
+      `;
-        OFFSET ${offset}
+      return rows.map(rowToSearchResult);
-      ),
+    } finally {
-      best_chunks AS (
+      await sql`SET statement_timeout = '0'`;
-        SELECT DISTINCT ON (rp.slug)
+    }
-          rp.slug, rp.id as page_id, rp.title, rp.type, rp.score,
-          cc.chunk_text, cc.chunk_source
-        FROM ranked_pages rp
-        JOIN content_chunks cc ON cc.page_id = rp.id
-        ORDER BY rp.slug, cc.chunk_index
-      )
-      SELECT slug, page_id, title, type, chunk_text, chunk_source, score,
-        false AS stale
-      FROM best_chunks
-      ORDER BY score DESC
-    `;
-
-    return rows.map(rowToSearchResult);
   }
 
   async searchVector(embedding: Float32Array, opts?: SearchOpts): Promise<SearchResult[]> {
@@ -231,23 +237,28 @@ export class PostgresEngine implements BrainEngine {
 
     const vecStr = '[' + Array.from(embedding).join(',') + ']';
 
-    const rows = await sql`
+    // Search-only timeout (see searchKeyword for rationale)
-      SELECT
+    await sql`SET statement_timeout = '8s'`;
-        p.slug, p.id as page_id, p.title, p.type,
+    try {
-        cc.chunk_text, cc.chunk_source,
+      const rows = await sql`
-        1 - (cc.embedding <=> ${vecStr}::vector) AS score,
+        SELECT
-        false AS stale
+          p.slug, p.id as page_id, p.title, p.type,
-      FROM content_chunks cc
+          cc.chunk_text, cc.chunk_source,
-      JOIN pages p ON p.id = cc.page_id
+          1 - (cc.embedding <=> ${vecStr}::vector) AS score,
-      WHERE cc.embedding IS NOT NULL
+          false AS stale
-        ${type ? sql`AND p.type = ${type}` : sql``}
+        FROM content_chunks cc
-        ${excludeSlugs?.length ? sql`AND p.slug != ALL(${excludeSlugs})` : sql``}
+        JOIN pages p ON p.id = cc.page_id
-      ORDER BY cc.embedding <=> ${vecStr}::vector
+        WHERE cc.embedding IS NOT NULL
-      LIMIT ${limit}
+          ${type ? sql`AND p.type = ${type}` : sql``}
-      OFFSET ${offset}
+          ${excludeSlugs?.length ? sql`AND p.slug != ALL(${excludeSlugs})` : sql``}
-    `;
+        ORDER BY cc.embedding <=> ${vecStr}::vector
-
+        LIMIT ${limit}
-    return rows.map(rowToSearchResult);
+        OFFSET ${offset}
+      `;
+      return rows.map(rowToSearchResult);
+    } finally {
+      await sql`SET statement_timeout = '0'`;
+    }
   }
 
   // Chunks

test/publish.test.ts

@@ -210,8 +210,9 @@ describe('generateHtml', () => {
     expect(html).toContain('prefers-color-scheme: dark');
   });
 
-  test('includes marked.js CDN', () => {
+  test('inlines marked.js (no CDN dependency)', () => {
     const html = generateHtml({ title: 'T', markdown: 'x' });
-    expect(html).toContain('cdn.jsdelivr.net/npm/marked');
+    expect(html).not.toContain('cdn.jsdelivr.net');
+    expect(html).toContain('marked');
   });
 });