GeezerVison

docs/contributing/repo-layout.md-7-31 (1)

7-31: ⚠️ Potential issue | 🟡 Minor | ⚡ Quick win

Label all fenced blocks with a language.

Lines 7, 37, 70, and 87 use unlabeled fences, which triggers MD040 repeatedly.

Suggested fix

-```
+```text
 bQuery/
 ...

```diff
-```
+```text
 src/
 ...

```diff
-```
+```text
 docs/
 ...

```diff
-```
+```text
 examples/
 ...

</details>
As per coding guidelines, "Run bun run lint before committing; follow ESLint flat config with `@typescript-eslint/recommended` and Prettier formatting."


Also applies to: 37-62, 70-83, 87-94

<details>
<summary>🤖 Prompt for AI Agents</summary>

Verify each finding against current code. Fix only still-valid issues, skip the
rest with a brief reason, keep changes minimal, and validate.

In @docs/contributing/repo-layout.md around lines 7 - 31, The Markdown file
contains unlabeled fenced code blocks showing directory trees (the bQuery/ ASCII
tree blocks and the similar src/, docs/, examples/ snippets) which triggers
MD040; update each of those backtick-fenced blocks to include a language label
(e.g., ```text) so the linter recognizes them, then reformat the file with the
repo's formatter and run the lint task (bun run lint) to ensure it complies with
the ESLint flat config and Prettier rules before committing.

</details>

</blockquote></details>
<details>
<summary>docs/contributing/architecture.md-7-9 (1)</summary><blockquote>

`7-9`: _⚠️ Potential issue_ | _🟡 Minor_ | _⚡ Quick win_

**Add a language tag to the fenced block.**

Line 7 uses an unlabeled fenced block, which triggers MD040 and can create docs-lint noise.

 

<details>
<summary>Suggested fix</summary>

```diff
-```
+```text
 view · store · router · forms · …  →  reactive  →  core
 ```
```
</details>
As per coding guidelines, "Run bun run lint before committing; follow ESLint flat config with `@typescript-eslint/recommended` and Prettier formatting."

<details>
<summary>🤖 Prompt for AI Agents</summary>

Verify each finding against current code. Fix only still-valid issues, skip the
rest with a brief reason, keep changes minimal, and validate.

In @docs/contributing/architecture.md around lines 7 - 9, The fenced code block
containing "view · store · router · forms · … → reactive → core" is
unlabeled and triggers MD040; update the block by adding a language tag (e.g.,
use ```text) to the opening fence so the block becomes a labeled fenced code
block, then run the docs linter (bun run lint) / Prettier to ensure formatting
follows the repo's ESLint/Prettier rules.

</details>

</blockquote></details>
<details>
<summary>docs/contributing/release-process.md-52-57 (1)</summary><blockquote>

`52-57`: _⚠️ Potential issue_ | _🟡 Minor_ | _⚡ Quick win_

**Specify the language for the commit-message code fence.**

Line 52 is an unlabeled fence and triggers MD040.

 

<details>
<summary>Suggested fix</summary>

```diff
-```
+```text
 feat(reactive): add useEventSource heartbeat option
 fix(server): validate cookie attributes for header-safe characters
 docs(ssr): document flushBoundary
 test(forms): cover disabled field early-return path
 ```
```
</details>
As per coding guidelines, "Run bun run lint before committing; follow ESLint flat config with `@typescript-eslint/recommended` and Prettier formatting."

<details>
<summary>🤖 Prompt for AI Agents</summary>

Verify each finding against current code. Fix only still-valid issues, skip the
rest with a brief reason, keep changes minimal, and validate.

In @docs/contributing/release-process.md around lines 52 - 57, The markdown code
fence containing the commit messages is missing a language identifier (triggers
MD040); update the unlabeled triple-backtick fence around the commit list in the
release-process document to use a language tag (e.g. text) so the fence becomes text and then close normally, ensuring the block exactly matches the
suggested fix; target the commit-message/code-fence block that contains lines
like "feat(reactive): add useEventSource heartbeat option" and "fix(server):
validate cookie attributes..." and run the repo linter (bun run lint) afterward
to confirm MD040 is resolved.

</details>

</blockquote></details>
<details>
<summary>docs/index.md-79-102 (1)</summary><blockquote>

`79-102`: _⚠️ Potential issue_ | _🟡 Minor_ | _⚡ Quick win_

**Add a language identifier to the fenced module tree block.**

This code fence is missing a language tag, which trips markdownlint (MD040).
 

<details>
<summary>Suggested fix</summary>

```diff
-```
+```text
 `@bquery/bquery`
 ├── core ─────────── selectors, traversal, events, utilities
 ...
 └── server ───────── dependency-free routing + WebSockets
 ```
```
</details>

<details>
<summary>🤖 Prompt for AI Agents</summary>

Verify each finding against current code. Fix only still-valid issues, skip the
rest with a brief reason, keep changes minimal, and validate.

In @docs/index.md around lines 79 - 102, The fenced module tree block starting
with "@bquery/bquery" in docs/index.md is missing a language identifier and
triggers markdownlint MD040; update the opening triple-backtick for that block
to include a language tag (e.g., ```text) so the tree block is treated as plain
text and linting passes, leaving the content (the lines listing core, reactive,
concurrency, etc.) unchanged.

</details>

</blockquote></details>
<details>
<summary>README.md-19-20 (1)</summary><blockquote>

`19-20`: _⚠️ Potential issue_ | _🟡 Minor_ | _⚡ Quick win_

**Fix blockquote spacing to satisfy markdownlint (MD028).**

There is a blank blockquote line here, which triggers `no-blanks-blockquote`. Keep the “New in 1.14.1” note as a single uninterrupted blockquote line.

<details>
<summary>Suggested fix</summary>

```diff
 > **New in 1.14.1:** `@bquery/bquery/motion` refreshes its cached reduced-motion media query whenever `window.matchMedia` changes, so `prefersReducedMotion()` and `reducedMotionSignal()` stay aligned with the active runtime implementation instead of returning stale values.
- 
```
</details>

<details>
<summary>🤖 Prompt for AI Agents</summary>

```
Verify each finding against current code. Fix only still-valid issues, skip the
rest with a brief reason, keep changes minimal, and validate.

In `@README.md` around lines 19 - 20, The blockquote that begins with "**New in
1.14.1:** `@bquery/bquery/motion` ..." contains an empty blank line which
triggers markdownlint MD028; remove the blank line so the note is one
uninterrupted blockquote line (i.e., collapse the separate ">" and the content
into a single "> **New in 1.14.1:** `@bquery/bquery/motion` refreshes..." block)
to satisfy the no-blanks-blockquote rule.
```

</details>

</blockquote></details>
<details>
<summary>docs/concepts/architecture.md-9-9 (1)</summary><blockquote>

`9-9`: _⚠️ Potential issue_ | _🟡 Minor_ | _⚡ Quick win_

**Add explicit language identifiers to fenced blocks.**

Line 9 and Line 68 use unlabeled fences, which triggers MD040 (`fenced-code-language`) and can cause docs-lint noise.

<details>
<summary>✅ Minimal fix</summary>

```diff
-```
+```text
             view · store · router · forms · …  (high-level)
                             │
                             ▼
                         reactive                (mid-level)
                             │
                             ▼
                           core                  (low-level)
 ```

```diff
-```
+```text
  ┌────────┐  user event  ┌─────────────┐  signal write  ┌──────────┐
  │  DOM   │ ───────────▶ │  handler/   │ ─────────────▶ │  signal  │
  │ event  │              │  store      │                │ / store  │
  └────────┘              └─────────────┘                └────┬─────┘
                                                              │ notify
                                                              ▼
                                                        ┌──────────┐
                                                        │ effects, │
                                                        │ computeds│
                                                        └────┬─────┘
                                                             │ update
                                                             ▼
                                                        ┌──────────┐
                                                        │   DOM    │
                                                        │ bindings │
                                                        └──────────┘
 ```
</details>





Also applies to: 68-68

<details>
<summary>🤖 Prompt for AI Agents</summary>

Verify each finding against current code. Fix only still-valid issues, skip the
rest with a brief reason, keep changes minimal, and validate.

In @docs/concepts/architecture.md at line 9, The two unlabeled fenced code
blocks that contain the ASCII diagrams should include an explicit language
identifier (e.g., ```text) to satisfy MD040; update the unlabeled
triple-backtick fences surrounding the ASCII diagrams (the fenced blocks that
show the "view · store · router · forms · …" diagram and the user-event →
handler/store → signal/store workflow diagram) by changing the opening backticks
to use a language tag like "text".

</details>

</blockquote></details>
<details>
<summary>docs/guide/api-core.md-407-407 (1)</summary><blockquote>

`407-407`: _⚠️ Potential issue_ | _🟡 Minor_ | _⚡ Quick win_

**Fix API name mismatch in pitfalls list.**

`cloneDeep` is referenced here, but this guide’s utility list documents `clone`. Please align the name to avoid confusion.

<details>
<summary>🤖 Prompt for AI Agents</summary>

```
Verify each finding against current code. Fix only still-valid issues, skip the
rest with a brief reason, keep changes minimal, and validate.

In `@docs/guide/api-core.md` at line 407, The pitfalls line currently mentions the
wrong utility name: replace the reference to cloneDeep with the documented
utility name clone so the sentence reads that utilities like deepEqual, clone,
and freeze operate by structure (symbols and non-enumerable props skipped);
update the text that references cloneDeep to clone wherever the guide’s utility
list uses clone to keep names consistent with the API (check occurrences around
the pitfalls list and any matching references to cloneDeep).
```

</details>

</blockquote></details>
<details>
<summary>docs/concepts/bundle-and-tree-shaking.md-29-33 (1)</summary><blockquote>

`29-33`: _⚠️ Potential issue_ | _🟡 Minor_ | _⚡ Quick win_

**Add a language tag to the fenced code block.**

This fence should be annotated (e.g., `text`) to satisfy markdown linting and keep formatting consistent.

<details>
<summary>Suggested fix</summary>

```diff
-```
+```text
 core · reactive · concurrency · component · motion · security · platform ·
 router · store · view · storybook · forms · i18n · a11y · dnd · media ·
 plugin · devtools · testing · ssr · server
 ```
```
</details>

<details>
<summary>🤖 Prompt for AI Agents</summary>

Verify each finding against current code. Fix only still-valid issues, skip the
rest with a brief reason, keep changes minimal, and validate.

In @docs/concepts/bundle-and-tree-shaking.md around lines 29 - 33, Update the
fenced code block that currently contains the list starting with "core ·
reactive · concurrency · component · motion · security · platform ·" to include
a language tag (e.g., text) after the opening ``` so the block becomes

closing ``` remains and no other content changes.

docs/guide/media.md-880-880 (1)

880-880: ⚠️ Potential issue | 🟡 Minor | ⚡ Quick win

Narrow this claim to avoid over-promising API support.

This line says all composables accept { signal: AbortSignal }, but the page earlier scopes that guarantee to the newly added composables in 1.14.0. Tightening wording here avoids misleading users on legacy APIs.

Suggested doc tweak

-- All composables accept `{ signal: AbortSignal }` (1.14.0) for auto-teardown — prefer it over manual `destroy()`.
+- New 1.14.0 composables accept `{ signal: AbortSignal }` for auto-teardown — prefer it where supported over manual `destroy()`.

🤖 Prompt for AI Agents

Verify each finding against current code. Fix only still-valid issues, skip the
rest with a brief reason, keep changes minimal, and validate.

In `@docs/guide/media.md` at line 880, The statement over-generalizes support for
`{ signal: AbortSignal }`; update the sentence to scope the guarantee to the new
composables introduced in 1.14.0 (e.g., "The newly added composables in 1.14.0
accept `{ signal: AbortSignal }` for auto-teardown — prefer it over manual
`destroy()`.") so readers don't assume legacy composables accept the option.

docs/guide/reactive.md-1164-1164 (1)

1164-1164: ⚠️ Potential issue | 🟡 Minor | ⚡ Quick win

EffectScope teardown method name is inconsistent.

This line uses scope.dispose(), but the same page documents EffectScope.stop() as the teardown API. Please align this to avoid copy/paste test failures.

🤖 Prompt for AI Agents

Verify each finding against current code. Fix only still-valid issues, skip the
rest with a brief reason, keep changes minimal, and validate.

In `@docs/guide/reactive.md` at line 1164, The doc uses scope.dispose() but the
API is EffectScope.stop(); update the guidance to use EffectScope.stop() (e.g.,
wrap setup in effectScope() and assert subscribers drop after scope.stop()) so
the symbols effectScope(), scope.stop(), and EffectScope.stop are consistent
with the documented API.

docs/guide/router.md-497-497 (1)

497-497: ⚠️ Potential issue | 🟡 Minor | ⚡ Quick win

beforeResolve is still navigation-blocking.

The guidance “Use beforeResolve for non-blocking work” is misleading—beforeResolve runs in the navigation pipeline too. Recommend rewording toward afterEach (or out-of-band side effects) for truly non-blocking work.

🤖 Prompt for AI Agents

Verify each finding against current code. Fix only still-valid issues, skip the
rest with a brief reason, keep changes minimal, and validate.

In `@docs/guide/router.md` at line 497, Update the sentence about navigation
guards to clarify that beforeResolve still runs in the navigation pipeline and
can block navigation; replace the phrase "Use `beforeResolve` for non-blocking
work" with guidance to use `afterEach` or perform out-of-band side effects for
truly non-blocking tasks, and ensure you mention `guards`, `beforeResolve`, and
`afterEach` so readers know the correct alternatives.

docs/guide/plugin.md-419-421 (1)

419-421: ⚠️ Potential issue | 🟡 Minor | ⚡ Quick win

Use plugin teardown API here, not store teardown API.

In this module’s testing section, recommending store clearPlugins() is misleading. For plugin registry cleanup, point to resetPlugins() from @bquery/bquery/plugin to keep teardown guidance consistent.

🤖 Prompt for AI Agents

Verify each finding against current code. Fix only still-valid issues, skip the
rest with a brief reason, keep changes minimal, and validate.

In `@docs/guide/plugin.md` around lines 419 - 421, The docs incorrectly recommend
using the store API clearPlugins() for plugin test teardown; update the guidance
to use the plugin registry teardown API resetPlugins() from
`@bquery/bquery/plugin` instead. Replace the mention of clearPlugins() with
resetPlugins() in the testing section and note that getInstalledPlugins({
withMetadata: true }) and getPluginInfo(name) should be paired with
resetPlugins() in afterEach to ensure proper plugin registry cleanup. Ensure the
text references the resetPlugins() symbol explicitly and removes or clarifies
any implication that clearPlugins() is the correct teardown for plugins.

docs/guide/server.md-422-422 (1)

422-422: ⚠️ Potential issue | 🟡 Minor | ⚡ Quick win

Sanitization bypass option is inconsistent in this page.

This line references { sanitize: false }, but earlier the guide instructs { trusted: true }. Please unify to the actual API option to avoid unsafe/copy-paste mistakes.

🤖 Prompt for AI Agents

Verify each finding against current code. Fix only still-valid issues, skip the
rest with a brief reason, keep changes minimal, and validate.

In `@docs/guide/server.md` at line 422, The docs mention two different options for
disabling HTML sanitization on ctx.html(): change the inconsistent `{ sanitize:
false }` usage to the actual API option `{ trusted: true }` (or vice‑versa to
match the library's real API) so the page consistently uses the same flag;
update every occurrence referencing ctx.html(...) options to the single correct
option name and ensure the explanatory note warns to only use it with fully
trusted content.

docs/cookbook/ssr-deploy.md-14-14 (1)

14-14: ⚠️ Potential issue | 🟡 Minor | ⚡ Quick win

Fix runtime count mismatch in the deployment note.

The table includes four targets (Node, Bun, Deno, Edge), but this line says “all three runtimes.” Please align wording to avoid confusion.

🤖 Prompt for AI Agents

Verify each finding against current code. Fix only still-valid issues, skip the
rest with a brief reason, keep changes minimal, and validate.

In `@docs/cookbook/ssr-deploy.md` at line 14, Update the sentence that reads "The
cross-runtime CI matrix (`.github/workflows/ssr-cross-runtime.yml`) guards the
public surface across all three runtimes." to correctly reflect the table's four
targets by changing "all three runtimes" to "all four runtimes" (or "all listed
runtimes"); locate the exact string in docs/cookbook/ssr-deploy.md and replace
it so the wording matches the Node, Bun, Deno, Edge targets.

docs/cookbook/spring-drag.md-11-11 (1)

11-11: ⚠️ Potential issue | 🟡 Minor | ⚡ Quick win

Avoid unsafe non-null casting in the cookbook snippet.

document.querySelector('.draggable') as HTMLElement can throw later if the element is absent. Prefer an explicit null guard so the example is safe to copy-paste.

Suggested doc-snippet fix

-const box = document.querySelector('.draggable') as HTMLElement;
+const box = document.querySelector<HTMLElement>('.draggable');
+if (!box) throw new Error('Missing .draggable element');

🤖 Prompt for AI Agents

Verify each finding against current code. Fix only still-valid issues, skip the
rest with a brief reason, keep changes minimal, and validate.

In `@docs/cookbook/spring-drag.md` at line 11, The snippet unsafely casts
document.querySelector('.draggable') as HTMLElement (variable "box"); change it
to use a typed query and an explicit null guard—query with
document.querySelector<HTMLElement>('.draggable') (or keep untyped), then check
if box is null and handle (early return or throw) before using it so the example
is safe to copy-paste.

docs/cookbook/theme-switcher.md-20-20 (1)

20-20: ⚠️ Potential issue | 🟡 Minor | ⚡ Quick win

Use a safe element lookup in the toggle handler snippet.

document.querySelector('#toggle')! can crash if the element is missing. Please show a guarded lookup pattern in this recipe.

Suggested doc-snippet fix

-document.querySelector('`#toggle`')!.addEventListener('click', () => {
-  stored.value = theme() === 'dark' ? 'light' : 'dark';
-});
+const toggle = document.querySelector<HTMLButtonElement>('`#toggle`');
+if (toggle) {
+  toggle.addEventListener('click', () => {
+    stored.value = theme() === 'dark' ? 'light' : 'dark';
+  });
+}

🤖 Prompt for AI Agents

Verify each finding against current code. Fix only still-valid issues, skip the
rest with a brief reason, keep changes minimal, and validate.

In `@docs/cookbook/theme-switcher.md` at line 20, The snippet uses
document.querySelector('`#toggle`')! which can throw if the element is missing;
change it to perform a safe lookup by assigning the result of
document.querySelector('`#toggle`') to a const (e.g., const toggle =
document.querySelector('`#toggle`')), check for null (if (!toggle) return or
handle accordingly), then call toggle.addEventListener('click', ...) — or use
optional chaining toggle?.addEventListener — and, if needed, narrow the type to
the appropriate HTMLElement before using element-specific properties.

docs/contributing/testing-strategy.md (1)

22-41: ⚡ Quick win

Avoid demonstrating raw innerHTML assignment in contributor examples.

Line 29 currently models a direct HTML write. Even in test snippets, this can conflict with the project’s security-by-default messaging; prefer DOM construction APIs in docs examples.

Suggested fix

-    root.innerHTML = '<button id="b">0</button>';
+    const button = document.createElement('button');
+    button.id = 'b';
+    button.textContent = '0';
+    root.appendChild(button);

As per coding guidelines, "Security: ALL DOM HTML writes MUST use `sanitizeHtml()` from `src/security/sanitize.ts`" and "Never write `innerHTML = untrustedInput`."

🤖 Prompt for AI Agents

Verify each finding against current code. Fix only still-valid issues, skip the
rest with a brief reason, keep changes minimal, and validate.

In `@docs/contributing/testing-strategy.md` around lines 22 - 41, The example
should avoid direct innerHTML assignment; replace root.innerHTML = '<button
id="b">0</button>' with safe DOM construction or a sanitized write: either
create a button via document.createElement('button'), set id and textContent
(e.g., const btn = document.createElement('button'); btn.id='b';
btn.textContent='0'; root.appendChild(btn)), or call sanitizeHtml(...) on the
string and use the sanitized result before assigning to root (import
sanitizeHtml from 'src/security/sanitize.ts'); update the test to reference the
constructed element (document.getElementById('b') or $('`#b`')) so signal, effect,
and $('`#b`') usage remains unchanged.

docs/.vitepress/config.ts (1)

3-5: ⚡ Quick win

Align constants with the repo camelCase rule.

SITE_URL and DESCRIPTION (Line 3 and Line 4) conflict with the TypeScript naming convention used in this repository for variables.

♻️ Suggested rename

-const SITE_URL = 'https://bquery.js.org';
-const DESCRIPTION =
+const siteUrl = 'https://bquery.js.org';
+const description =
   'Batteries-included TypeScript framework for the modern web — signals, SSR, Web Components, routing, and more — with a jQuery-inspired API and zero mandatory build step.';
...
-  description: DESCRIPTION,
+  description,
...
-    ['meta', { property: 'og:description', content: DESCRIPTION }],
-    ['meta', { property: 'og:url', content: SITE_URL }],
-    ['meta', { property: 'og:image', content: `${SITE_URL}/assets/bquerry-logo.svg` }],
+    ['meta', { property: 'og:description', content: description }],
+    ['meta', { property: 'og:url', content: siteUrl }],
+    ['meta', { property: 'og:image', content: `${siteUrl}/assets/bquerry-logo.svg` }],
...
-    ['meta', { name: 'twitter:description', content: DESCRIPTION }],
-    ['meta', { name: 'twitter:image', content: `${SITE_URL}/assets/bquerry-logo.svg` }],
+    ['meta', { name: 'twitter:description', content: description }],
+    ['meta', { name: 'twitter:image', content: `${siteUrl}/assets/bquerry-logo.svg` }],
...
-    hostname: SITE_URL,
+    hostname: siteUrl,

As per coding guidelines: **/*.ts: Keep file names kebab-case, classes PascalCase, and functions/variables camelCase.

Also applies to: 167-197

🤖 Prompt for AI Agents

Verify each finding against current code. Fix only still-valid issues, skip the
rest with a brief reason, keep changes minimal, and validate.

In `@docs/.vitepress/config.ts` around lines 3 - 5, The file declares top-level
constants SITE_URL and DESCRIPTION in SCREAMING_SNAKE_CASE which violates the
repo's camelCase rule; rename them to siteUrl and description (and any similar
constants in the file around lines 167-197) and update every usage within this
config (e.g., where SITE_URL and DESCRIPTION are referenced) to the new
camelCase identifiers (preserve values and exports) so the names follow the
project's camelCase variable convention.

docs/guide/a11y.md (1)

676-695: ⚡ Quick win

Align the test snippet with the repo’s DOM-test setup pattern.

This example uses document.body.innerHTML fixture injection; your test guidelines prefer creating elements inline and removing them after assertions. Updating the snippet would keep docs consistent with project practice.

As per coding guidelines: “Create elements inline, test functionality, then call .remove() to clean up; use happy-dom via tests/setup.ts for DOM simulation”.

🤖 Prompt for AI Agents

Verify each finding against current code. Fix only still-valid issues, skip the
rest with a brief reason, keep changes minimal, and validate.

In `@docs/guide/a11y.md` around lines 676 - 695, The test uses
document.body.innerHTML to inject a fixture; change it to create DOM elements
inline (e.g., createElement for main, h1, img, label, input), append the
container to document.body, call auditA11y() and assert on auditA11y().passed
and auditA11y().errors, then call .remove() on the container to clean up; locate
the snippet referencing auditA11y and document.body.innerHTML and replace with
element creation/append/remove flow so it matches the repo DOM-test setup
pattern.

docs/cookbook/notification-toast.md (1)

23-25: ⚡ Quick win

Use .peek() for one-time reduced-motion checks.

This snippet does a one-off signal read, so .peek() is the safer pattern for non-tracking access.

Suggested doc-snippet fix

-    if (!reducedMotion.value) {
+    if (!reducedMotion.peek()) {
       animate(this, { transform: ['translateY(20px)', 'translateY(0)'] }, { duration: 180 });
     }

As per coding guidelines: “Signal .value tracks dependencies → use .peek() to avoid tracking.”

🤖 Prompt for AI Agents

Verify each finding against current code. Fix only still-valid issues, skip the
rest with a brief reason, keep changes minimal, and validate.

In `@docs/cookbook/notification-toast.md` around lines 23 - 25, The one-time
reduced-motion check uses reducedMotion.value which creates tracking; change the
non-tracking read to use reducedMotion.peek() instead (i.e., call
prefersReducedMotion() to get reducedMotion and use reducedMotion.peek() in the
if condition) so the animate(...) call remains untracked; refer to the
prefersReducedMotion() call, the reducedMotion signal, and the animate(this,
...) usage when making this small change.

docs/workflows/forms-validation.md (1)

117-132: ⚡ Quick win

Use named export in the locale example to match repo TS conventions.

This snippet currently uses a default export, which conflicts with your TypeScript convention and can cause copy-paste lint friction.

Suggested doc snippet update

- export default {
+ export const en = {
   email: { label: 'Email' },
   password: { label: 'Password' },
   plan: { label: 'Plan' },
   submit: 'Create account',
   submitting: 'Working…',
   errors: {
     required: 'This field is required.',
     email: 'Enter a valid email address.',
     password: { short: 'Use at least 8 characters.' },
     plan: 'Choose a plan.',
   },
 };

As per coding guidelines, "**/*.ts: Use named exports only; no default exports".

🤖 Prompt for AI Agents

Verify each finding against current code. Fix only still-valid issues, skip the
rest with a brief reason, keep changes minimal, and validate.

In `@docs/workflows/forms-validation.md` around lines 117 - 132, The locale object
is exported as a default export but the repo requires named exports; change the
export to a named export (e.g., export const en = { email: {...}, password:
{...}, plan: {...}, submit: 'Create account', submitting: 'Working…', errors:
{...} }) and update any imports to use the named symbol (en) instead of the
default; update usages that import the locale to use the new named export to
avoid lint errors.

docs/workflows/streaming-ssr.md (1)

105-117: ⚡ Quick win

Prefer a named export in the edge handler example.

To keep docs aligned with your TS export conventions, avoid export default in this snippet.

Suggested doc snippet update

 import { createEdgeHandler, createSSRContext, renderToResponse } from '`@bquery/bquery/ssr`';
 
-export default createEdgeHandler(async (request) => {
+export const handleRequest = createEdgeHandler(async (request) => {
   const context = createSSRContext({ request, metrics });
   const data = {
     /* …resolve data from the request… */
   };
   return renderToResponse(compiled, data, {
     context,
     cache: { store: cache, vary: ['accept-language'] },
   });
 });

As per coding guidelines, "**/*.ts: Use named exports only; no default exports".

🤖 Prompt for AI Agents

Verify each finding against current code. Fix only still-valid issues, skip the
rest with a brief reason, keep changes minimal, and validate.

In `@docs/workflows/streaming-ssr.md` around lines 105 - 117, Replace the default
export with a named export in the edge handler example: instead of using "export
default createEdgeHandler(...)" declare a named export (e.g., "export const
handler = createEdgeHandler(...)") so the snippet follows the project's TS
convention; keep the existing use of createEdgeHandler, createSSRContext,
renderToResponse, compiled, cache and context variables intact and only change
the export form to a named export.

src/motion/reduced-motion.ts (1)

32-129: ⚡ Quick win

Add @internal JSDoc to the newly introduced internal helpers.

The new non-exported helpers should be annotated as internal to keep API intent explicit and guideline-compliant.

♻️ Suggested doc annotation pattern

+/**
+ * Safely resolves `window.matchMedia` and a bound callable.
+ * `@internal`
+ */
 const resolveMatchMedia = (): {
   readonly source: (query: string) => MediaQueryList;
   readonly call: (query: string) => MediaQueryList;
 } | null => {

+/**
+ * Re-syncs media-query subscription when the `matchMedia` source changes.
+ * `@internal`
+ */
 const syncMediaQuerySubscription = (

+/**
+ * Ensures the active reduced-motion media-query subscription exists.
+ * `@internal`
+ */
 const ensureMediaQuerySubscription = (

+/**
+ * Tears down the active reduced-motion media-query subscription.
+ * `@internal`
+ */
 const teardownMediaQuerySubscription = (): void => {

As per coding guidelines, src/**/*.ts: Mark non-public helpers with @internal JSDoc comment.

🤖 Prompt for AI Agents

Verify each finding against current code. Fix only still-valid issues, skip the
rest with a brief reason, keep changes minimal, and validate.

In `@src/motion/reduced-motion.ts` around lines 32 - 129, The new non-exported
helper functions (resolveMatchMedia, syncMediaQuerySubscription,
evaluateCurrent, dispatchIfChanged, ensureMediaQuerySubscription,
teardownMediaQuerySubscription) are missing `@internal` JSDoc annotations; add a
brief /** `@internal` */ JSDoc comment above each of these function declarations
to mark them as internal per the src/**/*.ts guidelines so the API intent is
explicit.

Fix all unresolved CodeRabbit comments on this PR:

• Push a commit to this branch (recommended)
• Create a new PR with the fixes