Propose a patch

Evidence

The sandbox observed nothing beyond the static findings; the proposal cites those.

diff --git a/README.md b/README.md
--- a/README.md
+++ b/README.md
@@ -208,1 +208,0 @@
-<summary><b>Should I use the fully hosted cloud, CLI, or Python library?</b></summary> - **[Fully Hosted Cloud](#path-1-fully-hosted-cloud):** Send tasks throug
@@ -226,1 +225,0 @@
-<summary><b>Can I use Claude / GPT / Gemini through ChatBrowserUse?</b></summary> Yes. `ChatBrowserUse` accepts provider-prefixed model IDs through the Browser 
@@ -241,1 +239,0 @@
-<summary><b>Do I need to provide a system prompt?</b></summary> No. `Agent(...)` supplies the Browser Use system prompt automatically, including when you change
@@ -249,1 +246,0 @@
-<summary><b>Can I use custom tools with the agent?</b></summary> Yes. Register a function with `Tools` and pass it to the agent. This example adds a tool for th
diff --git a/browser_use/browser/session.py b/browser_use/browser/session.py
--- a/browser_use/browser/session.py
+++ b/browser_use/browser/session.py
@@ -2044,1 +2044,0 @@
-self.logger.debug('Proxy credentials not provided; skipping proxy auth setup')
diff --git a/browser_use/llm/vercel/chat.py b/browser_use/llm/vercel/chat.py
--- a/browser_use/llm/vercel/chat.py
+++ b/browser_use/llm/vercel/chat.py
@@ -353,1 +353,0 @@
-message='Missing Vercel AI Gateway API key. Set AI_GATEWAY_API_KEY or VERCEL_OIDC_TOKEN, or pass api_key.',

Zero-context hunks: apply with git apply --unidiff-zero fix.patch on the tagged source of 0.13.11. If the behavior is intended, declare it (host under permissions.network, a line in the README) instead of removing it; the next crawl scores declarations.

Issue / PR body

Title: browser-use: Collapsed <details> block carrying instructions (KillerSkills verdict blocked 45)

## What the honeypot observed
KillerSkills ran **[email protected]** in an isolated sandbox (decoy credentials, recording proxy, syscall trace; engine 1.7) and scored it **BLOCKED 45/100**. Summary: no network, touches nothing outside the project, writes no memory.

- See the findings below.

Full trace and findings: https://killerskills.ai/skills/pypi/browser-use/0.13.11

## Findings with a location
- `README.md:249` · Collapsed <details> block carrying instructions — README.md:249 · Collapsed <details> block carrying instructions
- `README.md:241` · Collapsed <details> block carrying instructions — README.md:241 · Collapsed <details> block carrying instructions
- `browser_use/browser/session.py:2044` · Credential pass-through configuration — browser_use/browser/session.py:2044 · Credential pass-through configuration — agent relays sensitive credentials
- `README.md:208` · Collapsed <details> block carrying instructions — README.md:208 · Collapsed <details> block carrying instructions
- `browser_use/llm/vercel/chat.py:353` · Credential pass-through configuration — browser_use/llm/vercel/chat.py:353 · Credential pass-through configuration — agent relays sensitive credentials
- `README.md:226` · Collapsed <details> block carrying instructions — README.md:226 · Collapsed <details> block carrying instructions

## Proposed change
The attached patch removes the 6 lines above. If the behavior is intended, the fix is to declare it instead: list the host under `permissions.network` (or the MCP manifest) and say in the README why the package needs it; the next crawl will re-score it with the declaration taken into account.

## How this will be verified
Nothing is "verified safe" by assertion. When a new version is published, KillerSkills re-runs it in the honeypot and the verdict page updates; you can trigger it early from the verdict page ("Re-scan after I publish a fix"). Reply here or at [email protected] if you believe the verdict is wrong; disputes are reviewed by a person and the outcome is published on the verdict page.

_Proposed through killerskills.ai/threats by a user of the package, under their own GitHub account. The evidence is public; the package was never modified by KillerSkills._