chrome-automation
Connect to and control Google Chrome browser using agent-browser with CDP (Chrome DevTools Protocol). Use when the user wants to automate their existing Chrome browser, see browser actions in real-time, or needs to control the Chrome instance they're already using. Handles installation, setup, connecting via remote debugging, and all browser automation tasks with live visual feedback.
67
82%
Does it follow best practices?
Run evals on this skill
Adds up to 20 points to the overall score
View guide
Critical
Do not install without reviewing
Security
3 findings: 2 critical severity, 1 medium severity. Installing this skill is not recommended: please review these findings carefully if you do intend to do so.
E005: Suspicious download URL detected in skill instructions.
What this means
Detected a suspicious URL in the skill instructions that could lead the agent to download and execute malicious scripts or binaries. This includes links to executables from untrusted sources, typosquatting of official packages, URL shorteners that obscure the destination, and personal file hosting services.
Why it was flagged
The list is mostly official/benign (Node.js, Rust installers, Playwright, GitHub repo) but includes a direct raw GitHub shell script (https://raw.githubusercontent.com/vercel-labs/agent-browser/main/install.sh) that the package instructs users to curl|bash — a high-risk pattern that can deliver arbitrary code and should be treated as suspicious.
E006: Malicious code pattern detected in skill scripts.
What this means
Detected high-risk code patterns in the skill content — including its prompts, tool definitions, and resources — such as data exfiltration, backdoors, remote code execution, credential theft, system compromise, supply chain attacks, and obfuscation techniques.
Why it was flagged
The content contains deliberate behaviors that preserve and import a user's Chrome login state/cookies into a persistent automation profile and documents mechanisms to run remote install scripts and execute arbitrary JS in-browser, which can be used to access authenticated sessions and perform remote actions without explicit, ongoing user consent.
W013: Attempt to modify system services in skill instructions.
What this means
The skill prompts the agent to compromise the security or integrity of the user’s machine by modifying system-level services or configurations, such as obtaining elevated privileges, altering startup scripts, or changing system-wide settings.
Why it was flagged
This skill explicitly instructs running auto-install scripts that modify the host (clone repos, install Node/Rust, build native binaries), directs using PowerShell with ExecutionPolicy Bypass and running PowerShell as Administrator, and imports the user's Chrome profile (cookies/logins), all of which modify machine state and bypass security controls.
Low
Low-risk findings.
2 low severity findings. Worth noting, but not necessarily harmful.
W011: Third-party content exposure detected (indirect prompt injection risk).
What this means
The skill exposes the agent to untrusted, user-generated content from public third-party sources, creating a risk of indirect prompt injection. This includes browsing arbitrary URLs, reading social media posts or forum comments, and analyzing content from unknown websites.
Why it was flagged
SKILL.md mandates capturing and “read[ing] screenshot and describ[ing] to user” after browser actions; since the browser automation can navigate to arbitrary external pages and then extract/read their visible content, that visible page text can flow into the agent’s LLM context (indirectly) via the screenshot-reading/description step.
W012: Unverifiable external dependency detected (runtime URL that controls agent).
What this means
The skill fetches instructions or code from an external URL at runtime, and the fetched content directly controls the agent’s prompts or executes code. This dynamic dependency allows the external source to modify the agent’s behavior without any changes to the skill itself.
Why it was flagged
The skill's auto-install and install-scripts fetch and execute remote code (git clone of https://github.com/vercel-labs/agent-browser.git, curl | bash of https://raw.githubusercontent.com/vercel-labs/agent-browser/main/install.sh and rust installers from https://sh.rustup.rs and https://win.rustup.rs/x86_64) as part of the required setup, which clearly downloads and runs external code at runtime.
- Repository
- anbeime/skill
- Commit
f595e96
- Audited
- Security analysis
Is this your skill?
If you maintain this skill, you can claim it as your own. Once claimed, you can manage eval scenarios, bundle related skills, attach documentation or rules, and ensure cross-agent compatibility.