fix(extract): mirror plain-bullet Format 4 in parseTimelineEntries (db-source parity)

extractTimelineFromContent (fs-source) and parseTimelineEntries
(db-source extract + put_page auto_timeline) are documented as
kept-in-sync; adding Format 4 to only the fs parser left brain-authored
plain bullets invisible on the write path and `extract --source db`.
Same skip in the citation pass so a plain bullet carrying its own
[Source: ...] files exactly one entry on the db path too.

Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>

docs/tutorials/company-brain.md

@@ -223,16 +223,14 @@ export GBRAIN_REMOTE_CLIENT_ID=<Alice's client_id>
 export GBRAIN_REMOTE_CLIENT_SECRET=<Alice's client_secret>
 export GBRAIN_REMOTE_MCP_URL=https://brain.acme-co.com/mcp
 
-gbrain search "performance review"
+gbrain search "performance review" --remote
 ```
 
-(On a thin-client install every shared op routes through the remote MCP server automatically — no flag needed. The env vars select whose credentials the call uses.)
-
 Alice should see results only from `customers` and `shared`. The performance-review notes live in `internal`, which she's not scoped to read. She shouldn't see them.
 
 ```bash
 # Terminal 2, as Bob (export his credentials similarly)
-gbrain search "performance review"
+gbrain search "performance review" --remote
 ```
 
 Bob should see the performance-review notes from `internal`, plus anything related from `shared`. He shouldn't see anything that lives only in `customers`.
@@ -516,7 +514,7 @@ The first sync embeds every page, which takes time. Check `gbrain sources status
 
 ### "I see a page I shouldn't see"
 
-This shouldn't happen, but if you suspect it, run `gbrain search <query> --json` as the constrained client (thin-client install, with the client's `GBRAIN_REMOTE_*` env exported) and inspect the `source_id` field on every returned result. Every row should be in the client's `--federated-read` set. If one isn't, file an issue with the exact slug and source IDs.
+This shouldn't happen, but if you suspect it, run `gbrain search <query> --remote --json` as the constrained client and inspect the `source_id` field on every returned result. Every row should be in the client's `--federated-read` set. If one isn't, file an issue with the exact slug and source IDs.
 
 ### "The synthesized answer is wrong"
 

src/cli.ts

@@ -464,11 +464,7 @@ async function main() {
     // routed path. Date → ISO string; bigint → string (postgres.js shape);
     // Buffer → object. Microsecond-cost; eliminates a whole drift bug class.
     const result = JSON.parse(JSON.stringify(rawResult, bigintToStringReplacer));
-    // #380 pass-through: `--json` (undeclared on most ops, promised by docs)
-    // emits the raw op result instead of the human formatter.
-    const output = params.json === true
-      ? JSON.stringify(result, null, 2) + '\n'
-      : formatResult(op.name, result);
+    const output = formatResult(op.name, result);
     if (output) process.stdout.write(output);
   } catch (e: unknown) {
     // v0.42.20.0 (codex D4): on error, set exitCode + return so the `finally`
@@ -551,10 +547,7 @@ async function runThinClientRouted(
       signal: sigintController.signal,
     });
     const result = unpackToolResult(raw);
-    // #380: same --json seam as the local-engine path (renderer parity).
-    const output = params.json === true
-      ? JSON.stringify(result, null, 2) + '\n'
-      : formatResult(op.name, result);
+    const output = formatResult(op.name, result);
     if (output) process.stdout.write(output);
   } catch (e: unknown) {
     if (e instanceof RemoteMcpError) {
@@ -764,28 +757,10 @@ export function resolveQueryImage(
   return { path: imagePath, base64, mime };
 }
 
-/**
- * #380: undeclared flags that are honored DOWNSTREAM of parseOpArgs and must
- * keep passing through when unknown flags become hard errors:
- *   - source     → makeContext's resolveSourceId (the --source axis)
- *   - brain      → the mount/brain routing axis (docs promise the flag)
- *   - dry_run    → makeContext's ctx.dryRun (ops without a declared dry_run)
- *   - json       → raw-JSON output seam (local + thin-client paths)
- */
-const PASSTHROUGH_VALUE_FLAGS = new Set(['source', 'brain']);
-const PASSTHROUGH_BOOL_FLAGS = new Set(['dry_run', 'json']);
-
 export function parseOpArgs(op: Operation, args: string[]): Record<string, unknown> {
   const params: Record<string, unknown> = {};
   const positional = op.cliHints?.positional || [];
   let posIdx = 0;
-  const cliName = op.cliHints?.name || op.name;
-  const MAX_STDIN = 5_000_000; // 5MB cap, shared by stdin and --file
-  // #380: `--file <path>` fills the op's declared stdin param (put's `content`)
-  // from a file. Driven by cliHints.stdin — no per-op hard-coding — and
-  // disabled when the op declares a real `file` param of its own.
-  const fileParam = op.cliHints?.stdin && !op.params.file ? op.cliHints.stdin : undefined;
-  let filePath: string | undefined;
 
   for (let i = 0; i < args.length; i++) {
     const arg = args[i];
@@ -799,42 +774,12 @@ export function parseOpArgs(op: Operation, args: string[]): Record<string, unkno
         }
       }
       const key = arg.slice(2).replace(/-/g, '_');
-      if (fileParam && key === 'file') {
-        if (i + 1 >= args.length) {
-          console.error(`Error: ${arg} requires a value.`);
-          process.exit(1);
-        }
-        filePath = args[++i];
-        continue;
-      }
       const paramDef = op.params[key];
-      if (!paramDef) {
-        if (PASSTHROUGH_BOOL_FLAGS.has(key)) {
-          params[key] = true;
-          continue;
-        }
-        if (PASSTHROUGH_VALUE_FLAGS.has(key)) {
-          if (i + 1 >= args.length) {
-            console.error(`Error: ${arg} requires a value.`);
-            process.exit(1);
-          }
-          params[key] = args[++i];
-          continue;
-        }
-        // #380: unknown flags were silently swallowed into params, so typos
-        // like `put --file` created empty pages instead of erroring.
-        console.error(`Unknown option for gbrain ${cliName}: ${arg}`);
-        console.error(`Run 'gbrain ${cliName} --help' for valid flags.`);
-        process.exit(1);
-      }
-      if (paramDef.type === 'boolean') {
+      if (paramDef?.type === 'boolean') {
         params[key] = true;
       } else if (i + 1 < args.length) {
         params[key] = args[++i];
-        if (paramDef.type === 'number') params[key] = Number(params[key]);
-      } else {
-        console.error(`Error: ${arg} requires a value.`);
-        process.exit(1);
+        if (paramDef?.type === 'number') params[key] = Number(params[key]);
       }
     } else if (posIdx < positional.length) {
       const key = positional[posIdx++];
@@ -843,30 +788,10 @@ export function parseOpArgs(op: Operation, args: string[]): Record<string, unkno
     }
   }
 
-  // #380: resolve --file AFTER the loop so --file/--content conflicts are
-  // caught in either order.
-  if (filePath !== undefined && fileParam) {
-    if (params[fileParam] !== undefined) {
-      console.error(`Error: use only one of --file, --${fileParam}, or stdin for gbrain ${cliName}.`);
-      process.exit(1);
-    }
-    let fileContent: string;
-    try {
-      fileContent = readFileSync(filePath, 'utf-8');
-    } catch (e) {
-      console.error(`Error: cannot read --file ${filePath}: ${e instanceof Error ? e.message : String(e)}`);
-      process.exit(1);
-    }
-    if (Buffer.byteLength(fileContent, 'utf-8') > MAX_STDIN) {
-      console.error(`Error: file content exceeds ${MAX_STDIN} bytes. Split into smaller inputs.`);
-      process.exit(1);
-    }
-    params[fileParam] = fileContent;
-  }
-
   // Read stdin for content params
   if (op.cliHints?.stdin && !params[op.cliHints.stdin] && !process.stdin.isTTY) {
     const stdinContent = readFileSync(0, 'utf-8');
+    const MAX_STDIN = 5_000_000; // 5MB
     if (Buffer.byteLength(stdinContent, 'utf-8') > MAX_STDIN) {
       console.error(`Error: stdin content exceeds ${MAX_STDIN} bytes. Split into smaller inputs.`);
       process.exit(1);
@@ -2328,10 +2253,6 @@ export function printOpHelp(op: Operation, invokedName?: string) {
       const prefix = isPos ? `  <${key}>` : `  --${key.replace(/_/g, '-')}`;
       console.log(`${prefix.padEnd(28)} ${def.description || ''}${req}`);
     }
-    // #380: ops that read stdin also accept --file <path> (parseOpArgs).
-    if (op.cliHints?.stdin && !op.params.file) {
-      console.log(`${'  --file <path>'.padEnd(28)} Read ${op.cliHints.stdin} from a file (alternative to --${op.cliHints.stdin} or stdin)`);
-    }
   }
 }
 

src/commands/extract.ts

@@ -509,8 +509,12 @@ export function extractTimelineFromContent(content: string, slug: string): Extra
   // DB-level uniqueness cannot collapse.
   const citationPattern = /\[Source:\s*([^\]]+?),\s*(\d{4}-\d{2}-\d{2})\s*\]/g;
   const bulletLinePattern = /^-\s+\*\*\d{4}-\d{2}-\d{2}\*\*\s*\|/;
+  // Lines captured by Format 4 (plain bullet) are skipped for the same
+  // reason: quality.md mandates a trailing [Source: ...] on those bullets,
+  // and re-extracting the citation would double-count the event.
+  const plainBulletLinePattern = /^-\s+\d{4}-\d{2}-\d{2}\s*[—–-]/;
   for (const line of content.split(/\r?\n/)) {
-    if (bulletLinePattern.test(line)) continue;
+    if (bulletLinePattern.test(line) || plainBulletLinePattern.test(line)) continue;
     const lineMatches = [...line.matchAll(citationPattern)];
     if (lineMatches.length === 0) continue;
     // Strip every citation marker from the line to leave the annotated text.
@@ -526,6 +530,19 @@ export function extractTimelineFromContent(content: string, slug: string): Extra
     }
   }
 
+  // Format 4: Plain bullet — - YYYY-MM-DD — Summary
+  // This is the format gbrain's own enrich skill writes (no bold, no source
+  // pipe). Without it, every brain-authored timeline entry is invisible to
+  // extraction and timeline_coverage stays at 0%. Anchored at line start so a
+  // date inside a summary/link cannot start a spurious entry; a bold Format-1
+  // line (`- **…**`) cannot match here (a `*` follows the bullet, not a
+  // digit), and the citation loop above skips these lines, so a plain bullet
+  // carrying its own [Source: ...] files exactly one entry.
+  const plainBulletPattern = /^-\s+(\d{4}-\d{2}-\d{2})\s*[—–-]\s*(.+)$/gm;
+  while ((match = plainBulletPattern.exec(content)) !== null) {
+    entries.push({ slug, date: match[1], source: 'markdown', summary: match[2].trim() });
+  }
+
   return entries;
 }
 

src/core/link-extraction.ts

@@ -1103,6 +1103,11 @@ export interface TimelineCandidate {
 // Match: `- **YYYY-MM-DD** | summary` or `- **YYYY-MM-DD** -- summary`
 // or `- **YYYY-MM-DD** - summary` or just `**YYYY-MM-DD** | summary`.
 const TIMELINE_LINE_RE = /^\s*-?\s*\*\*(\d{4}-\d{2}-\d{2})\*\*\s*[|\-–—]+\s*(.+?)\s*$/;
+// Plain bullet: `- YYYY-MM-DD — summary` (no bold, no source pipe). Kept in
+// sync with extractTimelineFromContent's Format 4 (the fs-source path).
+// Anchored flush-left so a date inside an indented continuation line cannot
+// start a spurious entry.
+const PLAIN_TIMELINE_LINE_RE = /^-\s+(\d{4}-\d{2}-\d{2})\s*[—–-]\s*(.+?)\s*$/;
 
 /**
  * Parse timeline entries from content. Looks at:
@@ -1119,7 +1124,7 @@ export function parseTimelineEntries(content: string): TimelineCandidate[] {
 
   let i = 0;
   while (i < lines.length) {
-    const m = TIMELINE_LINE_RE.exec(lines[i]);
+    const m = TIMELINE_LINE_RE.exec(lines[i]) ?? PLAIN_TIMELINE_LINE_RE.exec(lines[i]);
     if (!m) {
       i++;
       continue;
@@ -1136,7 +1141,7 @@ export function parseTimelineEntries(content: string): TimelineCandidate[] {
     let j = i + 1;
     while (j < lines.length) {
       const next = lines[j];
-      if (TIMELINE_LINE_RE.test(next)) break;
+      if (TIMELINE_LINE_RE.test(next) || PLAIN_TIMELINE_LINE_RE.test(next)) break;
       if (/^#{1,6}\s/.test(next)) break;
       if (next.trim().length === 0 && detailLines.length === 0) {
         // skip leading blank line; if we hit a blank after detail content
@@ -1166,7 +1171,7 @@ export function parseTimelineEntries(content: string): TimelineCandidate[] {
   // bullet pass are skipped (a bullet often carries its own citation).
   const citationRe = /\[Source:\s*([^\]]+?),\s*(\d{4}-\d{2}-\d{2})\s*\]/g;
   for (const line of lines) {
-    if (TIMELINE_LINE_RE.test(line)) continue;
+    if (TIMELINE_LINE_RE.test(line) || PLAIN_TIMELINE_LINE_RE.test(line)) continue;
     const matches = [...line.matchAll(citationRe)];
     if (matches.length === 0) continue;
     const summary = line

src/core/operations.ts

@@ -769,7 +769,7 @@ const get_page: Operation = {
 
 const put_page: Operation = {
   name: 'put_page',
-  description: 'Write/update a page (markdown with frontmatter). Chunks, embeds, reconciles tags, and (when auto_link/auto_timeline are enabled) extracts + reconciles graph links and timeline entries. On the CLI, `gbrain put SLUG --file PATH` reads content from a file (also `--content` or stdin). For provenance write-through and a binary-NUL guard, prefer `gbrain capture --file PATH --slug SLUG` (v0.39.3.0).',
+  description: 'Write/update a page (markdown with frontmatter). Chunks, embeds, reconciles tags, and (when auto_link/auto_timeline are enabled) extracts + reconciles graph links and timeline entries. For large content on Windows (pipe-buffer limit ~45KB) or any file-as-input workflow, use `gbrain capture --file PATH --slug SLUG` — capture reads the file as a Buffer with a binary-NUL guard and adds provenance write-through (v0.39.3.0).',
   params: {
     slug: { type: 'string', required: true, description: 'Page slug' },
     content: { type: 'string', required: true, description: 'Full markdown content with YAML frontmatter' },
@@ -1384,10 +1384,7 @@ const list_pages: Operation = {
   params: {
     type: { type: 'string', description: 'Filter by page type' },
     tag: { type: 'string', description: 'Filter by tag' },
-    limit: { type: 'number', description: 'Max results (default 50, capped at 100 — use offset to paginate beyond)' },
-    // #2876: the 100-row cap was silent and there was no way past it even
-    // though both engines already support OFFSET on listPages.
-    offset: { type: 'number', description: 'Skip first N results (pagination; pair with limit)' },
+    limit: { type: 'number', description: 'Max results (default 50)' },
     // v0.29 — surface filter that already exists on PageFilters.
     updated_after: {
       type: 'string',
@@ -1418,10 +1415,6 @@ const list_pages: Operation = {
       type: p.type as any,
       tag: p.tag as string,
       limit: clampSearchLimit(p.limit as number | undefined, 50, 100),
-      // #2876: thread pagination through (engines already honor offset).
-      offset: Number.isFinite(p.offset as number) && (p.offset as number) > 0
-        ? Math.floor(p.offset as number)
-        : undefined,
       includeDeleted: (p.include_deleted as boolean) === true,
       updated_after: typeof p.updated_after === 'string' ? p.updated_after : undefined,
       sort,

test/cli-args.test.ts

@@ -1,41 +1,8 @@
 import { describe, expect, test } from 'bun:test';
-import { mkdtempSync, rmSync, writeFileSync } from 'fs';
-import { tmpdir } from 'os';
-import { join } from 'path';
 import { parseOpArgs } from '../src/cli.ts';
 import { operationsByName } from '../src/core/operations.ts';
 
 describe('parseOpArgs', () => {
-  // #380: `gbrain put SLUG --file PATH` reads content from the file instead
-  // of silently swallowing the flag and creating an empty page.
-  test('put --file reads the stdin param (content) from a file', () => {
-    const dir = mkdtempSync(join(tmpdir(), 'gbrain-put-file-'));
-    try {
-      const pagePath = join(dir, 'page.md');
-      writeFileSync(pagePath, '# From file\n\nBody loaded from --file.\n');
-      const params = parseOpArgs(operationsByName.put_page, ['concepts/from-file', '--file', pagePath]);
-      expect(params.slug).toBe('concepts/from-file');
-      expect(params.content).toBe('# From file\n\nBody loaded from --file.\n');
-    } finally {
-      rmSync(dir, { recursive: true, force: true });
-    }
-  });
-
-  // #380 regression guard: undeclared-but-honored flags must keep passing
-  // through when unknown flags become hard errors (--source is read by
-  // makeContext; --json by the output seam; --dry-run by ctx.dryRun).
-  test('pass-through allowlist flags survive on ops that do not declare them', () => {
-    const params = parseOpArgs(operationsByName.get_page, [
-      'people/alice-example', '--source', 'wiki', '--json', '--dry-run',
-    ]);
-    expect(params).toEqual({
-      slug: 'people/alice-example',
-      source: 'wiki',
-      json: true,
-      dry_run: true,
-    });
-  });
-
   test('--no-<boolean> maps to false without consuming the next flag', () => {
     const params = parseOpArgs(operationsByName.query, [
       'freshEmbedSourceScope code source',

test/cli.test.ts

@@ -1,5 +1,5 @@
 import { describe, test, expect } from 'bun:test';
-import { existsSync, mkdtempSync, readFileSync, rmSync, writeFileSync } from 'fs';
+import { existsSync, mkdtempSync, readFileSync, rmSync } from 'fs';
 import { tmpdir } from 'os';
 import { join } from 'path';
 
@@ -120,76 +120,6 @@ describe('CLI dispatch integration', () => {
     expect(exitCode).toBe(0);
   });
 
-  // #380 / PR #856: put --help documents the --file input path.
-  test('put --help documents --file input', async () => {
-    const proc = Bun.spawn(['bun', 'run', 'src/cli.ts', 'put', '--help'], {
-      cwd: repoRoot,
-      stdout: 'pipe',
-      stderr: 'pipe',
-    });
-    const stdout = await new Response(proc.stdout).text();
-    const exitCode = await proc.exited;
-    expect(stdout).toContain('Usage: gbrain put');
-    expect(stdout).toContain('--file <path>');
-    expect(exitCode).toBe(0);
-  });
-
-  // #380: unknown flags on shared ops are a hard error (previously silently
-  // swallowed into params — `put --file` created empty pages). parseOpArgs
-  // runs BEFORE engine connect, so the error must fire without a brain.
-  test('unknown shared-op flags fail before DB connection', async () => {
-    const home = mkdtempSync(join(tmpdir(), 'gbrain-cli-unknown-flag-'));
-    try {
-      const proc = Bun.spawn(['bun', 'run', 'src/cli.ts', 'get', 'people/alice', '--bogus'], {
-        cwd: repoRoot,
-        stdout: 'pipe',
-        stderr: 'pipe',
-        env: isolatedEnv(home),
-      });
-      const stderr = await new Response(proc.stderr).text();
-      const exitCode = await proc.exited;
-      expect(stderr).toContain('Unknown option for gbrain get: --bogus');
-      expect(stderr).not.toContain('No brain configured');
-      expect(exitCode).toBe(1);
-    } finally {
-      rmSync(home, { recursive: true, force: true });
-    }
-  });
-
-  test('put rejects combining --file and --content', async () => {
-    const home = mkdtempSync(join(tmpdir(), 'gbrain-cli-put-conflict-'));
-    try {
-      const pagePath = join(home, 'page.md');
-      writeFileSync(pagePath, 'file body\n');
-      const proc = Bun.spawn(
-        ['bun', 'run', 'src/cli.ts', 'put', 'a/b', '--content', 'inline', '--file', pagePath],
-        { cwd: repoRoot, stdout: 'pipe', stderr: 'pipe', env: isolatedEnv(home) },
-      );
-      const stderr = await new Response(proc.stderr).text();
-      const exitCode = await proc.exited;
-      expect(stderr).toContain('use only one of --file, --content, or stdin');
-      expect(exitCode).toBe(1);
-    } finally {
-      rmSync(home, { recursive: true, force: true });
-    }
-  });
-
-  test('put --file with a missing path errors instead of writing an empty page', async () => {
-    const home = mkdtempSync(join(tmpdir(), 'gbrain-cli-put-missing-file-'));
-    try {
-      const proc = Bun.spawn(
-        ['bun', 'run', 'src/cli.ts', 'put', 'a/b', '--file', join(home, 'nope.md')],
-        { cwd: repoRoot, stdout: 'pipe', stderr: 'pipe', env: isolatedEnv(home) },
-      );
-      const stderr = await new Response(proc.stderr).text();
-      const exitCode = await proc.exited;
-      expect(stderr).toContain('cannot read --file');
-      expect(exitCode).toBe(1);
-    } finally {
-      rmSync(home, { recursive: true, force: true });
-    }
-  });
-
   test('upgrade --help prints usage without running upgrade', async () => {
     const proc = Bun.spawn(['bun', 'run', 'src/cli.ts', 'upgrade', '--help'], {
       cwd: repoRoot,

test/extract.test.ts

@@ -185,6 +185,49 @@ describe('extractTimelineFromContent', () => {
     expect(entries).toHaveLength(1);
     expect(entries[0].summary).toBe('Landed the enterprise pilot with acme-example.');
   });
+
+  // Format 4: the plain bullet `- YYYY-MM-DD — Summary` that gbrain's own
+  // enrich skill writes. Before this was supported, every brain-authored
+  // timeline entry was invisible and timeline_coverage was stuck at 0%.
+  it('extracts plain bullet format (- YYYY-MM-DD — Summary)', () => {
+    const content = `## Timeline\n- 2026-06-01 — Catch-up call with alice-example; intros offered.`;
+    const entries = extractTimelineFromContent(content, 'people/alice-example');
+    expect(entries).toHaveLength(1);
+    expect(entries[0].date).toBe('2026-06-01');
+    expect(entries[0].source).toBe('markdown');
+    expect(entries[0].summary).toBe('Catch-up call with alice-example; intros offered.');
+  });
+
+  it('files exactly one entry for a plain bullet that carries its own citation', () => {
+    // quality.md mandates this shape; it must not double-count via the
+    // citation arm (Format 3) AND the plain-bullet arm (Format 4).
+    const content = `- 2026-05-31 — Confirmed full name. [Source: User, 2026-05-31]`;
+    const entries = extractTimelineFromContent(content, 'people/charlie-example');
+    expect(entries).toHaveLength(1);
+    expect(entries[0].date).toBe('2026-05-31');
+    expect(entries[0].source).toBe('markdown');
+    expect(entries[0].summary).toContain('Confirmed full name.');
+  });
+
+  it('does not double-count bold (Format 1) lines as plain bullets', () => {
+    const content = `- **2025-03-18** | Meeting — Discussed partnership`;
+    const entries = extractTimelineFromContent(content, 'test');
+    expect(entries).toHaveLength(1);
+    expect(entries[0].source).toBe('Meeting');
+  });
+
+  it('does not start a spurious entry from a date inside a summary', () => {
+    const content = `- 2026-06-01 — See [meeting](meetings/2026-06-01).`;
+    const entries = extractTimelineFromContent(content, 'people/alice-example');
+    expect(entries).toHaveLength(1);
+    expect(entries[0].date).toBe('2026-06-01');
+  });
+
+  it('handles en dash and hyphen separators in plain bullets', () => {
+    const content = `- 2026-06-01 – First\n- 2026-06-02 - Second`;
+    const entries = extractTimelineFromContent(content, 'test');
+    expect(entries).toHaveLength(2);
+  });
 });
 
 describe('walkMarkdownFiles', () => {

test/link-extraction.test.ts

@@ -720,6 +720,30 @@ More prose here.
     const entries = parseTimelineEntries(content);
     expect(entries.length).toBe(2);
   });
+
+  // Plain bullet (extract.ts Format 4 parity — the db-source path must see
+  // the same entries as the fs-source path).
+  test('parses plain bullet: - YYYY-MM-DD — summary', () => {
+    const entries = parseTimelineEntries('## Timeline\n- 2026-06-01 — Catch-up call with alice-example');
+    expect(entries.length).toBe(1);
+    expect(entries[0].date).toBe('2026-06-01');
+    expect(entries[0].summary).toBe('Catch-up call with alice-example');
+  });
+
+  test('files exactly one entry for a plain bullet carrying its own citation', () => {
+    const entries = parseTimelineEntries('- 2026-05-31 — Confirmed full name. [Source: User, 2026-05-31]');
+    expect(entries.length).toBe(1);
+    expect(entries[0].date).toBe('2026-05-31');
+  });
+
+  test('skips invalid dates in plain bullets', () => {
+    expect(parseTimelineEntries('- 2026-13-45 — Bad date').length).toBe(0);
+  });
+
+  test('does not double-count a bold bullet as a plain bullet', () => {
+    const entries = parseTimelineEntries('- **2026-01-15** | Met with Alice');
+    expect(entries.length).toBe(1);
+  });
 });
 
 // ─── isAutoLinkEnabled ─────────────────────────────────────────

test/list-pages-offset.test.ts

@@ -1,55 +0,0 @@
-import { describe, test, expect } from 'bun:test';
-import { operationsByName } from '../src/core/operations.ts';
-
-/**
- * #2876: `gbrain list --limit` silently clamped at 100 with no pagination.
- * list_pages now declares `offset` (both engines already supported it on
- * PageFilters) and the limit description discloses the 100-row cap.
- */
-describe('list_pages pagination (#2876)', () => {
-  const listPagesOp = operationsByName.list_pages;
-
-  function makeCtx(captured: unknown[]) {
-    return {
-      engine: {
-        listPages: async (filters: unknown) => {
-          captured.push(filters);
-          return [];
-        },
-      },
-      config: { engine: 'pglite' },
-      logger: { info() {}, warn() {}, error() {} },
-      dryRun: false,
-      remote: false,
-      sourceId: 'default',
-    } as any;
-  }
-
-  test('declares offset param and discloses the 100-row cap on limit', () => {
-    expect(listPagesOp.params.offset).toBeDefined();
-    expect(listPagesOp.params.offset.type).toBe('number');
-    expect(listPagesOp.params.limit.description).toContain('100');
-  });
-
-  test('threads offset through to engine.listPages', async () => {
-    const captured: any[] = [];
-    await listPagesOp.handler(makeCtx(captured), { limit: 10, offset: 30 });
-    expect(captured[0].offset).toBe(30);
-    expect(captured[0].limit).toBe(10);
-  });
-
-  test('drops negative, non-finite, and zero offsets', async () => {
-    const captured: any[] = [];
-    const ctx = makeCtx(captured);
-    await listPagesOp.handler(ctx, { offset: -5 });
-    await listPagesOp.handler(ctx, { offset: Infinity });
-    await listPagesOp.handler(ctx, { offset: 0 });
-    for (const f of captured) expect(f.offset).toBeUndefined();
-  });
-
-  test('floors fractional offsets', async () => {
-    const captured: any[] = [];
-    await listPagesOp.handler(makeCtx(captured), { offset: 7.9 });
-    expect(captured[0].offset).toBe(7);
-  });
-});