@remcp/runtime 0.2.75 → 0.2.77

package/CHANGELOG.md

@@ -2,6 +2,18 @@
 
 ## Unreleased
 
+## 0.2.77 — 2026-09-21
+
+- Fixed `browser_action` typing so the documented `text` argument reaches CDP `Input.insertText` with Unicode intact; `value` and `text_value` remain compatible aliases.
+- Clarified the browser action schema so agents use `text` for typing and reusable selectors from `browser_find`.
+- Lockstep patch release with the ReMCP client/server at `0.2.77`.
+
+## 0.2.76 — 2026-09-21
+
+- `browser_find(text=...)` ranks exact semantic targets ahead of broad ancestor text containers, prefers actionable/semantic nodes, and de-duplicates nested matches before applying the result limit.
+- Native `h1`–`h6` elements derive the `heading` role, improving browser text targeting and follow-up actions.
+- Lockstep patch release with the ReMCP client/server at `0.2.76`.
+
 ## 0.2.63 — 2026-09-20
 
 - Lockstep MCP Apps production-surface release with the ReMCP client/server at `0.2.63`; device runtime tool behavior is unchanged from `0.2.62`.

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "@remcp/runtime",
-  "version": "0.2.75",
+  "version": "0.2.77",
   "description": "First-party ReMCP local device runtime: filesystem, terminal, computer-use, browser, diagnostics and document tools over MCP.",
   "type": "module",
   "license": "MIT",

package/src/extended/browser.mjs

@@ -447,7 +447,7 @@ export async function browserSnapshot(args) {
   });
 }
 
-function findExpression(args) {
+export function findExpression(args) {
   const selector = optionalString(args.selector);
   const needle = optionalString(args.text);
   const role = optionalString(args.role);
@@ -474,6 +474,7 @@ function findExpression(args) {
       const explicit = (el.getAttribute('role') || '').trim().toLowerCase();
       if (explicit) return explicit;
       const tag = el.tagName.toLowerCase();
+      if (/^h[1-6]$/.test(tag)) return 'heading';
       if (tag === 'button') return 'button';
       if (tag === 'a' && el.hasAttribute('href')) return 'link';
       if (tag === 'textarea') return 'textbox';
@@ -495,15 +496,57 @@ function findExpression(args) {
       }
       return '';
     }
-    return source.filter(el => {
+    function visibleText(el) {
+      return (el.innerText || el.textContent || el.getAttribute('aria-label') || '').trim();
+    }
+    function textQuality(value) {
+      if (!needle) return 0;
+      const hay = value.toLowerCase();
+      const wanted = needle.toLowerCase();
+      if (hay === wanted) return 0;
+      if (hay.startsWith(wanted)) return 1;
+      return 2;
+    }
+    function depthOf(el) {
+      let depth = 0;
+      for (let current = el.parentElement; current; current = current.parentElement) depth += 1;
+      return depth;
+    }
+    const candidates = source.filter(el => {
       const style = getComputedStyle(el);
       const visible = style.visibility !== 'hidden' && style.display !== 'none' && el.getClientRects().length > 0;
-      const hay = (el.innerText || el.textContent || el.getAttribute('aria-label') || '').trim();
+      const hay = visibleText(el);
       const r = semanticRole(el);
       return visible && (!needle || hay.toLowerCase().includes(needle.toLowerCase())) && (!role || r === role.toLowerCase());
-    }).slice(0, ${limit}).map(el => {
+    }).map((el, index) => ({
+      el,
+      index,
+      text: visibleText(el),
+      role: semanticRole(el),
+      depth: depthOf(el),
+    }));
+    let selected = candidates;
+    if (needle && !selector) {
+      selected = [...candidates].sort((a, b) =>
+        textQuality(a.text) - textQuality(b.text)
+        || Number(!a.role) - Number(!b.role)
+        || a.text.length - b.text.length
+        || b.depth - a.depth
+        || a.index - b.index
+      );
+      const deduped = [];
+      for (const candidate of selected) {
+        if (deduped.some(existing => existing.el.contains(candidate.el) || candidate.el.contains(existing.el))) continue;
+        deduped.push(candidate);
+        if (deduped.length >= ${limit}) break;
+      }
+      selected = deduped;
+    } else {
+      selected = selected.slice(0, ${limit});
+    }
+    return selected.map(({ el }) => {
       const b = el.getBoundingClientRect();
-      return { selector: cssPath(el), tag: el.tagName.toLowerCase(), role: semanticRole(el), name: el.getAttribute('aria-label') || el.getAttribute('name') || '', text: (el.innerText || el.textContent || '').trim().slice(0, 500), value: 'value' in el ? String(el.value).slice(0, 500) : '', x: b.x, y: b.y, width: b.width, height: b.height, disabled: Boolean(el.disabled) };
+      return { selector: cssPath(el), tag: el.tagName.toLowerCase(), role: semanticRole(el), name: el.getAttribute('aria-label') || el.getAttribute('name') || '', text: visibleText(el).slice(0, 500), value: 'value' in el ? String(el.value).slice(0, 500) : '', x: b.x, y: b.y, width: b.width, height: b.height, disabled: Boolean(el.disabled) };
     });
   })()`;
 }
@@ -529,6 +572,12 @@ function elementLookup(args) {
   })()`;
 }
 
+export function browserActionInputValue(args = {}, action = '') {
+  // text is the public field models naturally use for type, while set_value/select historically
+  // used value/text_value. Keep all aliases compatible but give each action its semantic field.
+  return String(action === 'type' ? (args.text ?? args.text_value ?? args.value ?? '') : (args.value ?? args.text_value ?? args.text ?? ''));
+}
+
 export async function browserAction(args) {
   const action = requireEnum(args.action, 'action', ['click', 'focus', 'type', 'set_value', 'select', 'scroll_into_view', 'upload', 'press', 'set_viewport']);
   return withTarget(args, async (session, target) => {
@@ -568,7 +617,7 @@ export async function browserAction(args) {
       return jsonResult({ target_id: target.id, action, key });
     }
     const lookup = elementLookup(args);
-    const value = String(args.value ?? args.text_value ?? '');
+    const value = browserActionInputValue(args, action);
     const option = String(args.option ?? value);
 
     if (action === 'click') {

package/src/extended/catalog.mjs

@@ -555,7 +555,7 @@ export const extendedToolDefinitions = [
   define('browser_navigate', 'Browser navigate', 'Open a new debuggable tab or navigate a selected browser page by URL, history back/forward, or reload through Chrome DevTools Protocol.', o({ ...browserTarget, action:e(['url','new_tab','back','forward','reload'], 'Default url when url is provided, otherwise reload. new_tab creates the first page target when needed.'), url:s('Destination URL for action=url or action=new_tab.'), wait:b('Wait for the page load event; default true.'), ignore_cache:b('Reload without cache when action=reload.') }), openMutatingNonDestructive, browserHandlers.browser_navigate, ['browser_cdp']),
   define('browser_snapshot', 'Browser snapshot', 'Return the selected browser page accessibility tree through Chrome DevTools Protocol and, when requested, attach a real rendered viewport PNG for visual verification. A selector is temporarily recentered for capture, its exact viewport bounds are returned, and the prior scroll position is restored before return.', o({ ...browserTarget, max_nodes:n('Maximum AX nodes; default 1500.'), include_screenshot:b('Attach a real CDP-rendered viewport PNG screenshot; default false.'), selector:s('When include_screenshot=true, center this CSS-selected element before capture and report its viewport bounds.') }), readOnlyLive, browserHandlers.browser_snapshot, ['browser_cdp']),
   define('browser_find', 'Find browser element', 'Find visible DOM elements by CSS selector, text, or ARIA role and return reusable selectors plus text, value, and bounding boxes.', o({ ...browserTarget, selector:s('CSS selector.'), text:s('Visible text substring.'), role:s('ARIA role.'), limit:n() }), readOnlyLive, browserHandlers.browser_find, ['browser_cdp']),
-  define('browser_action', 'Browser element action', 'Click, focus, type/set a value, select, scroll to, upload files, press a key, or set an exact responsive-test viewport in a debuggable browser page using DOM/CDP semantics.', o({ ...browserTarget, action:e(['click','focus','type','set_value','select','scroll_into_view','upload','press','set_viewport']), selector:s(), text:s(), value:s(), text_value:s(), option:s(), path:s(), paths:{type:'array',items:{type:'string'}}, key:s(), width:n('CSS viewport width for set_viewport.'), height:n('CSS viewport height for set_viewport.'), device_scale_factor:n('Device scale factor for set_viewport; default 1.'), mobile:b('Enable mobile emulation for set_viewport.') }, ['action']), openMutating, browserHandlers.browser_action, ['browser_cdp']),
+  define('browser_action', 'Browser element action', 'Click, focus, type/set a value, select, scroll to, upload files, press a key, or set an exact responsive-test viewport in a debuggable browser page using DOM/CDP semantics.', o({ ...browserTarget, action:e(['click','focus','type','set_value','select','scroll_into_view','upload','press','set_viewport']), selector:s('CSS selector for the target element; prefer the reusable selector returned by browser_find.'), text:s('For action=type, Unicode text to insert. For element actions without selector, visible text may identify the target.'), value:s('Value for action=set_value; also accepted as a compatibility fallback for action=type.'), text_value:s('Compatibility alias for the type/set_value text value.'), option:s(), path:s(), paths:{type:'array',items:{type:'string'}}, key:s(), width:n('CSS viewport width for set_viewport.'), height:n('CSS viewport height for set_viewport.'), device_scale_factor:n('Device scale factor for set_viewport; default 1.'), mobile:b('Enable mobile emulation for set_viewport.') }, ['action']), openMutating, browserHandlers.browser_action, ['browser_cdp']),
   define('browser_wait', 'Wait for browser', 'Wait for DOM state, page text, URL, completed loading, a navigation away from the current URL, or a short network-idle period in a debuggable browser page. Use browser_evaluate separately for JavaScript predicates or page-specific inspection.', o({ ...browserTarget, condition:e(['selector','text','url_contains','load','navigation','network_idle']), selector:s(), text:s(), value:s(), poll_ms:n(), idle_ms:n('Required zero-in-flight network quiet window for network_idle; default 500 ms.') }), readOnlyLive, browserHandlers.browser_wait, ['browser_cdp']),
   define('browser_evaluate', 'Evaluate browser JavaScript', 'Evaluate JavaScript in a selected debuggable browser page and return its serializable value. This is a powerful escape hatch for page-specific automation.', o({ ...browserTarget, expression:s('JavaScript expression.'), await_promise:b('Await a returned Promise; default true.') }, ['expression']), openMutating, browserHandlers.browser_evaluate, ['browser_cdp']),