diff --git a/tools/benchmarks/results/phase1-markdown-template-picker-lazy-boundary-2026-07-30.md b/tools/benchmarks/results/phase1-markdown-template-picker-lazy-boundary-2026-07-30.md
new file mode 100644
index 00000000000..f4371ec271c
--- /dev/null
+++ b/tools/benchmarks/results/phase1-markdown-template-picker-lazy-boundary-2026-07-30.md
@@ -0,0 +1,153 @@
+# Phase 1 Markdown template picker lazy-boundary audit — 2026-07-30
+
+**Result:** rejected. The measured lazy-on-first-request candidate reduced the main-window
+static closure by 1,144 raw bytes, but automatic shared-chunk allocation added one JavaScript
+file and raw bytes to both the dashboard-popout and web closures. Only this negative report
+is retained; the candidate implementation, tests, and output were removed, the accepted
+source/output baseline was restored exactly, and no budget changed.
+
+## Production ownership and behavior audit
+
+The accepted source path is:
+
+`App.tsx` → `MarkdownTemplatePicker.tsx` → `CommandDialog`
+
+`MarkdownTemplatePicker` owns the only eager `subscribeMarkdownTemplatePicker` listener and
+keeps the active request in both state and a ref. A replacement request cancels the prior
+request before installing the new identity. Unmount cancels the active request and removes
+the singleton listener. Cancel, blank, and template selections all clear the active identity
+before invoking its once-wrapped resolver.
+
+The request module intentionally is not a durable store:
+
+- a request with zero templates resolves immediately to `{ type: 'blank' }`;
+- a request made without the mounted listener also resolves immediately to blank; and
+- a delivered request has one generated ID, the original template identities, and a resolver
+  protected by `once`.
+
+`createUntitledMarkdownFileWithTemplateSelection` discovers templates in the selected runtime
+and requests a selection before creating the file. The runtime operation arguments continue
+to carry the selected local, WSL, SSH, relay, or remote-runtime identity; the picker itself
+does not assume a local filesystem or git worktree. Folder workspaces and provider/Git policy
+are outside this UI request boundary and were not changed.
+
+The candidate kept a narrow eager `MarkdownTemplatePickerHost` at the exact existing
+`RecoverableRenderErrorBoundary` placement in `App.tsx`. The host retained listener,
+replacement, unmount, and resolution ownership; `lazyWithRetry` dynamically imported only
+the existing visual surface with reload key `markdown-template-picker` after the first real
+request. A permanent mounted latch kept the loaded surface mounted through later close and
+reopen cycles, while `Suspense` used the existing blank fallback. Focused candidate tests
+proved no pre-request load, zero-template/no-listener blank behavior, replacement while the
+chunk was unresolved, unmount cancellation, once-only resolution, selection identity, retry
+key, and mounted-instance continuity.
+
+## Why the candidate was rejected
+
+B emitted the intended dynamic surface:
+
+- `src/components/editor/MarkdownTemplatePicker.tsx` →
+  `assets/MarkdownTemplatePicker-BhSHOf-f.js`
+- 3,510 raw / 997 gzip bytes
+- SHA-256 `2324892a6e93f33e40761f32129df5a4d5dc1f54b700afb662ae7df00a6150ed`
+- dynamic importer `_App-BMTSSdCQ.js`
+
+The surface was absent from all three B static entry closures. However, its `CommandDialog`
+dependencies caused Rolldown to extract `assets/utils-CpTw4IZK.js` as a new shared static
+chunk: 63,446 raw / 13,707 gzip bytes, SHA-256
+`de0bc0dc34d80e5522f95e3e1e0e0d6bd5f76105c028e5246d988919f44a06fa`.
+That allocation added one static JavaScript file to every renderer entry through these exact
+B manifest paths:
+
+- `index.html` → `_checkbox-D-XjhT6S.js` → `_utils-CpTw4IZK.js`
+- `popout.html` → `_dropdown-menu-BC06w94O.js` → `_utils-CpTw4IZK.js`
+- `web-index.html` → `_label-DgUD63C_.js` → `_utils-CpTw4IZK.js`
+
+The split was beneficial only to the main-window raw total. Because both unrelated renderer
+entries grew in raw bytes and JavaScript count, B failed the strict retention rule. Reusing a
+misleading settings-owned loader or widening this into shared UI/chunk partitioning was
+outside the requested narrow ownership boundary and was not attempted.
+
+## Fresh A/B production evidence
+
+The paths below were ephemeral local build directories used during measurement, not durable
+artifacts. The hashes, byte counts, manifests, and conclusions recorded here are the portable
+evidence.
+
+- A artifact: `/tmp/orca-markdown-template-picker-a.uAneXM`
+- Rejected B artifact: `/tmp/orca-markdown-template-picker-b.yphkkd`
+- A transformed 2,003 main, 17 preload, and 9,186 renderer modules.
+- B transformed 2,003 main, 17 preload, and 9,187 renderer modules.
+- Both builds emitted only the two existing CSS `::highlight(...)` parser warnings.
+
+| Static closure   |     A raw |     B raw | Raw change |    A gzip |    B gzip | Gzip change | A JS | B JS | A CSS | B CSS |
+| ---------------- | --------: | --------: | ---------: | --------: | --------: | ----------: | ---: | ---: | ----: | ----: |
+| Main window      | 8,416,540 | 8,415,396 |     -1,144 | 1,877,823 | 1,878,138 |        +315 |  292 |  293 |     2 |     2 |
+| Dashboard popout | 4,507,253 | 4,507,716 |       +463 |   984,615 |   985,062 |        +447 |   77 |   78 |     2 |     2 |
+| Web renderer     | 4,360,652 | 4,361,076 |       +424 |   928,355 |   928,758 |        +403 |   33 |   34 |     1 |     1 |
+
+Electron entries were file-for-file identical:
+
+| Entry   | A/B raw | A/B gzip | A/B SHA-256                                                        |
+| ------- | ------: | -------: | ------------------------------------------------------------------ |
+| Main    | 776,873 |  174,092 | `6e75b9d4862f1219c23f16d6c920004167d7dae1995498caaacc5354a9d8f8fd` |
+| Preload | 130,798 |   20,642 | `c388a39cdca9609760e286d95b87ad1e53793720450e507f830ff1f6c5bd259f` |
+
+The complete A and B main trees each contain 184 files and have identical sorted SHA-256
+manifests:
+`3576805e0f10c1c6c3ca473257901f824326f4e2b1a0a224bbb50e72eb28a5f2`.
+The preload trees each contain one file and have identical sorted SHA-256 manifests:
+`3bb30bdb361c7c99cc423e4a4939399f8cb29042d653bdbfe5ef582034d9ed00`.
+
+The renderer manifests contain 778 A entries and 780 B entries. Their file SHA-256 values are
+`92ad617f8c8aee6d4c4a23622f524266cbe9c93b3fe65544d1acb833dc3f0d67` and
+`113190ab27844d76cde6e001933ed2aadfe50bcb1bee6449b3ed06dc3fd63c3b`,
+respectively. The complete output trees contain 972 A files and 974 B files; their sorted
+SHA-256 manifests are
+`09ff908615c9ec1f7f45bfe4f8929e56c2afa25c4a2b32242665d184f4b15df0` and
+`8f1b49d7f289874113a4329e6d81e86e4693315d30e97980ef44465a5777bb41`.
+
+Sorted static-closure manifests include path, raw, gzip, and file SHA-256:
+
+| Entry  | A rows / manifest SHA-256                                                | B rows / manifest SHA-256                                                |
+| ------ | ------------------------------------------------------------------------ | ------------------------------------------------------------------------ |
+| Main   | 294 / `f863c014e40dd4cd49b0864e4fa61c5df98f185712f76b9227fbb5d2cc443488` | 295 / `55f29097cf79973a7b579fba9ae68b50ed7202232fe380855e0086b3fc4ff1a5` |
+| Popout | 79 / `b83774dc8005e78455d394f6342824c9c21784c07813e43a3770d6c19ab61d45`  | 80 / `df349c8b2ada54e6f852dca9ff585f5d62c45498dbeaff1c0611caa3cfbe844c`  |
+| Web    | 34 / `46925643199b3b8f53da06bed01b5bed8ed070d79fb8b6214d70cf796672e571`  | 35 / `10d5726430fe7176e08ea501089fbcf89677c9bd5bc3621e2360801d7194ec16`  |
+
+The accepted renderer-index raw budget remains 8,466,347, preserving its 49,807-byte
+headroom over A. Electron main remains within its 825,109 raw budget with exactly 48,236
+bytes of headroom. No budget was changed.
+
+## Restoration and validation
+
+After rejecting B, only the candidate source and test edits were removed and the accepted
+source was rebuilt. The restored build transformed 2,003 main, 17 preload, and 9,186 renderer
+modules and is file-for-file identical to A across all 972 output files. Its sorted output
+manifest SHA-256 is
+`09ff908615c9ec1f7f45bfe4f8929e56c2afa25c4a2b32242665d184f4b15df0`.
+
+All 778 A and 780 B renderer manifest entries were checked. Every import key,
+dynamic-import key, emitted JavaScript/CSS file, and declared asset exists beneath its
+renderer artifact root; both arms had zero missing-target or escaping-path failures.
+
+- Candidate-focused Vitest plus `create-untitled-markdown.test.ts`: 2 files and 11 tests
+  passed before the rejected candidate tests were removed.
+- Accepted focused `create-untitled-markdown.test.ts`: 1 file and 7 tests passed.
+- `pnpm run typecheck:web`: passed.
+- Targeted `oxlint --deny-warnings`: passed over the picker, request, creation, App, and
+  focused test source.
+- Targeted `oxfmt --check`: passed over the same source plus this report.
+- `pnpm run check:max-lines-ratchet`: passed with 354 grandfathered suppressions and no new
+  bypasses.
+- `pnpm run check:electron-bundle-budgets`: passed against the restored A output.
+- `git diff --check`: passed.
+
+## Residual risk
+
+The accepted picker remains eager. A future attempt needs an allocation strategy that keeps
+the picker visual dynamic without extracting a new shared static utility chunk into popout
+and web; that is a bundler/shared-UI concern rather than a safe change to request ownership.
+
+No packaged launch smoke was run on macOS, Linux, or Windows. Production builds and manifest
+validation ran on this macOS worktree, so packaged cross-platform launch behavior remains
+unresolved.