fix: trim CODEOWNERS reviewers and stabilize search benchmark

Author: clean6378-max-it

.github/CODEOWNERS

@@ -1,20 +1,18 @@
 # Code owners — auto-requested as reviewers on matching paths.
 # Last match wins; see https://docs.github.com/en/repositories/managing-your-repositorys-settings-and-features/customizing-your-repository/about-code-owners
 #
-# Roles:
-#   Development + code review: @clean6378-max-it (Chen), @timon0305 (Zilin)
-#   Final approval + merge:    @wpak-ai (Will Pak)
+# Review owners (development + code review): @clean6378-max-it (Chen), @timon0305 (Zilin)
 #
-# PR authors do not self-merge. Get a code review from Chen or Zilin, then final
-# approval from Will before merge.
+# Final approval, merge sequencing, and no-self-merge rules are enforced by branch
+# protection / automation, not by listing additional owners here.
 
 # Default owners for the repository.
-*                       @clean6378-max-it @timon0305 @wpak-ai
+*                       @clean6378-max-it @timon0305
 
 # Docs and contributor process
-/docs/                  @clean6378-max-it @timon0305 @wpak-ai
-/CONTRIBUTING.md        @clean6378-max-it @timon0305 @wpak-ai
-/README.md              @clean6378-max-it @timon0305 @wpak-ai
+/docs/                  @clean6378-max-it @timon0305
+/CONTRIBUTING.md        @clean6378-max-it @timon0305
+/README.md              @clean6378-max-it @timon0305
 
 # CI and workflow changes
-/.github/               @clean6378-max-it @timon0305 @wpak-ai
+/.github/               @clean6378-max-it @timon0305

README.md

@@ -145,7 +145,7 @@ claude-code-chat-browser/
 
 See **[`CONTRIBUTING.md`](CONTRIBUTING.md)** for full setup, conventions, and where to change each layer. New contributors should follow **[`docs/onboarding.md`](docs/onboarding.md)** for a first-PR walkthrough and reading order.
 
-**Maintainer coverage:** commit activity is currently concentrated on a small set of identities (bus-factor risk). Mitigations: [`.github/CODEOWNERS`](.github/CODEOWNERS) splits development/code review (@clean6378-max-it, @timon0305) from final approval and merge (@wpak-ai), plus [`docs/onboarding.md`](docs/onboarding.md) for contributor ramp-up.
+**Maintainer coverage:** commit activity is currently concentrated on a small set of identities (bus-factor risk). Mitigations: [`.github/CODEOWNERS`](.github/CODEOWNERS) routes code review to @clean6378-max-it and @timon0305; final merge approval (@wpak-ai) is enforced via branch protection, plus [`docs/onboarding.md`](docs/onboarding.md) for contributor ramp-up.
 
 Quick start:
 

docs/onboarding.md

@@ -125,7 +125,7 @@ Recent commit history is concentrated on a small set of identities. If those mai
 
 **Mitigations in this repo:**
 
-- **[`.github/CODEOWNERS`](../.github/CODEOWNERS)** — @clean6378-max-it and @timon0305 for development and code review; @wpak-ai for final approval and merge.
+- **[`.github/CODEOWNERS`](../.github/CODEOWNERS)** — @clean6378-max-it and @timon0305 for code review routing. Final merge approval (@wpak-ai) is enforced via branch protection.
 - **This onboarding path** — lowers the ramp for additional contributors to run gates and ship safely.
 
 If you are joining as a reviewer, read the [suggested reading order](#suggested-reading-order) and run the [full local gate](#5-run-the-full-local-gate) once on `master` before approving your first PR.

tests/benchmarks/conftest.py

@@ -51,13 +51,14 @@ def tracemalloc_peak() -> TracemallocPeak:
 def write_jsonl(path: Path, line_count: int, *, first_timestamp: str | None = None) -> Path:
     """Write a JSONL session file with *line_count* rows derived from the template fixture."""
     template = json.loads(TEMPLATE_LINE)
+    if first_timestamp is not None:
+        base_dt = datetime.fromisoformat(first_timestamp.replace("Z", "+00:00"))
+    else:
+        base_dt = _EXPORT_SESSION_BASE.replace(hour=10)
     with path.open("w", encoding="utf-8") as f:
         for i in range(line_count):
             entry = deepcopy(template)
-            if i == 0 and first_timestamp is not None:
-                entry["timestamp"] = first_timestamp
-            else:
-                entry["timestamp"] = f"2026-06-12T10:{i % 60:02d}:00Z"
+            entry["timestamp"] = (base_dt + timedelta(minutes=i)).strftime("%Y-%m-%dT%H:%M:%SZ")
             if i % 3 == 1:
                 msg = entry.setdefault("message", {})
                 if isinstance(msg, dict) and "content" in msg:
@@ -79,8 +80,11 @@ def seed_search_corpus(
     """Create a multi-session project tree under *base_dir* for search benchmarks."""
     project = base_dir / "bench-project"
     project.mkdir(parents=True, exist_ok=True)
+    # Keep every message inside the default 30-day search window (see DEFAULT_SEARCH_WINDOW_DAYS).
+    anchor = datetime.now(UTC) - timedelta(days=1)
     for i in range(session_count):
-        write_jsonl(project / f"session_{i:04d}.jsonl", lines_per_session)
+        first_ts = (anchor + timedelta(minutes=i * 30)).strftime("%Y-%m-%dT%H:%M:%SZ")
+        write_jsonl(project / f"session_{i:04d}.jsonl", lines_per_session, first_timestamp=first_ts)
     return base_dir