Security Audit

Agent Browser

github.com/openclaw/skills

AI SkillCommit 9c1b8e80763e

CRITICAL

Scanned about 2 months ago

Critical

Immediate action required

High

Priority fixes suggested

Medium

Best practices review

Low

Acknowledged / Tracked

Trust Assessment

Agent Browser received a trust score of 40/100, placing it in the Untrusted category. This skill has significant security findings that require attention before use in production.

SkillShield's automated analysis identified 2 findings: 2 critical, 0 high, 0 medium, and 0 low severity. Key findings include Obfuscated execution: base64 decode to shell, Malicious Base64-encoded script execution during MacOS installation.

The analysis covered 4 layers: Manifest Analysis, Static Code Analysis, Dependency Graph, LLM Behavioral Safety. All layers scored 70 or above, reflecting consistent security practices.

Last analyzed on February 12, 2026 (commit 9c1b8e80). SkillShield performs automated 4-layer security analysis on AI skills and MCP servers.

Layer Breakdown

Manifest Analysis

100%

Static Code Analysis

70%

Dependency Graph

100%

LLM Behavioral Safety

70%

Behavioral Risk Signals

Network Access

1 finding

Shell Execution

2 findings

Dynamic Code

1 finding

Security Findings2

Severity	Finding	Layer	Location
CRITICAL	Obfuscated execution: base64 decode to shell Detected base64-encoded data being decoded and piped to a shell interpreter. This is a common obfuscation technique. Decode and inspect the base64 content before execution.	Static	skills/sakaen736jih/agent-browser-6aigix9qi2tu/SKILL.md:10
CRITICAL	Malicious Base64-encoded script execution during MacOS installation The MacOS installation instructions for 'OpenClawProvider' include a command that decodes a base64 string and pipes it directly to `bash`. The decoded string executes a `curl` command to download and execute a script from a highly suspicious, non-standard IP address (`http://91.92.242.30`). This constitutes a severe supply chain risk and command injection vulnerability, as it allows arbitrary code execution from an external, untrusted source during the skill's setup process, potentially compromising the user's system. The `Installer-Package` echo appears to be a decoy. Immediately remove the malicious MacOS installation command. Provide a legitimate and secure installation method for OpenClawProvider, or clarify that OpenClawProvider is a separate, potentially risky dependency not directly managed by the `agent-browser` skill. If OpenClawProvider is truly required, its installation should be from a trusted, verifiable source, and not involve obfuscated, externally hosted scripts.	LLM	SKILL.md:14

Scan History

Embed Code

[![SkillShield](https://skillshield.io/api/v1/badge/048173eab14997f4.svg)](https://skillshield.io/report/048173eab14997f4)