diff --git a/browser_use/skills/browser-use/SKILL.md b/browser_use/skills/browser-use/SKILL.md
index 0c8f6cc220..356e12fe98 100644
--- a/browser_use/skills/browser-use/SKILL.md
+++ b/browser_use/skills/browser-use/SKILL.md
@@ -44,6 +44,10 @@ PY
 - Invoke as `browser-use`. Use heredocs for multi-line commands.
 - Helpers are pre-imported. `run.py` calls `ensure_daemon()` before `exec`.
 - First navigation is `new_tab(url)`, not `goto_url(url)`.
+- `new_tab()` and `switch_tab()` attach and move the horse marker without
+  changing Chrome's visible tab. Screenshots and normal CDP input work in the
+  background; call `activate_tab(target)` only when the user explicitly asks
+  or a page demonstrably pauses rendering while hidden.
 - The normal local flow attaches to the running Chrome/Chromium CDP endpoint. No browser ids or local profile selection.
 
 ## Local Chrome
diff --git a/pyproject.toml b/pyproject.toml
index c3c3073a83..ba89a26adf 100644
--- a/pyproject.toml
+++ b/pyproject.toml
@@ -2,7 +2,7 @@
 name = "browser-use"
 description = "Make websites accessible for AI agents"
 authors = [{ name = "Gregor Zunic" }]
-version = "0.13.7"
+version = "0.13.8"
 readme = "README.md"
 requires-python = ">=3.11,<4.0"
 classifiers = [
@@ -41,12 +41,12 @@ dependencies = [
     "reportlab==4.4.9",
     "cdp-use==1.4.5",
     "pyotp==2.9.0",
-    "pillow==12.2.0",
+    "pillow==12.3.0",
     "cloudpickle==3.1.2",
     "markdownify==1.2.2",
     "python-docx==1.2.0",
     "browser-use-sdk==3.4.2",
-    "browser-harness==0.1.8",
+    "browser-harness==0.1.9",
 ]
 # google-api-core: only used for Google LLM APIs
 # pyperclip: only used for examples that use copy/paste
@@ -60,11 +60,11 @@ dependencies = [
 [project.optional-dependencies]
 cli = []
 core = [
-    "browser-use-core==0.13.2; sys_platform == 'darwin' and platform_machine == 'arm64'",
-    "browser-use-core==0.13.2; sys_platform == 'darwin' and platform_machine == 'x86_64'",
-    "browser-use-core==0.13.2; sys_platform == 'linux' and platform_machine == 'x86_64'",
-    "browser-use-core==0.13.2; sys_platform == 'linux' and platform_machine == 'aarch64'",
-    "browser-use-core==0.13.2; sys_platform == 'win32' and (platform_machine == 'AMD64' or platform_machine == 'x86_64')",
+    "browser-use-core==0.13.3; sys_platform == 'darwin' and platform_machine == 'arm64'",
+    "browser-use-core==0.13.3; sys_platform == 'darwin' and platform_machine == 'x86_64'",
+    "browser-use-core==0.13.3; sys_platform == 'linux' and platform_machine == 'x86_64'",
+    "browser-use-core==0.13.3; sys_platform == 'linux' and platform_machine == 'aarch64'",
+    "browser-use-core==0.13.3; sys_platform == 'win32' and (platform_machine == 'AMD64' or platform_machine == 'x86_64')",
 ]
 aws = ["boto3==1.42.37"]
 oci = ["oci==2.166.0"]
diff --git a/skills/browser-use/SKILL.md b/skills/browser-use/SKILL.md
index 0c8f6cc220..356e12fe98 100644
--- a/skills/browser-use/SKILL.md
+++ b/skills/browser-use/SKILL.md
@@ -44,6 +44,10 @@ PY
 - Invoke as `browser-use`. Use heredocs for multi-line commands.
 - Helpers are pre-imported. `run.py` calls `ensure_daemon()` before `exec`.
 - First navigation is `new_tab(url)`, not `goto_url(url)`.
+- `new_tab()` and `switch_tab()` attach and move the horse marker without
+  changing Chrome's visible tab. Screenshots and normal CDP input work in the
+  background; call `activate_tab(target)` only when the user explicitly asks
+  or a page demonstrably pauses rendering while hidden.
 - The normal local flow attaches to the running Chrome/Chromium CDP endpoint. No browser ids or local profile selection.
 
 ## Local Chrome