Extension

AI Security Review

Evidence context: threat category none; evidence quality moderate.

This extension's security findings are driven by well-documented false positive patterns rather than actual malicious behavior. The 1329 IoC findings are almost entirely XIOC extractor garbage—property access chains in JavaScript being misread as domain names. For example, fm.finally.call, i.open, q.events, self.location.search, e.page, and n.stream are not real domains but JavaScript property access chains that the IoC extractor incorrectly parsed as network indicators. The single legitimate-looking domain zloirock.ru is associated with the Brotli compression library, a legitimate open-source project.

The network findings show only 3 total detections, with one being a standard fetch call in next/static/chunks/377-3ce641142b1c8377.js—a Next.js bundled file, which is expected for modern web applications. Critically, there are zero malware signatures detected, zero obfuscation findings, and zero code-smell findings. This is the opposite of what we'd expect from malicious extensions, which typically combine obfuscation with actual malware signatures.

The extension's anonymous status (empty name and developer name fields) is concerning, but anonymity alone does not constitute evidence of malicious behavior. The user count of 2 is extremely low, suggesting this is either a very new or very obscure extension. However, the findings themselves reveal nothing harmful—the code does not access credentials, does not hijack browsers, does not contain malware signatures, and does not exhibit proxyware behavior.

Counterargument: A skeptic might argue that the anonymous publisher combined with 1333 findings should trigger a higher-risk verdict. However, the nature of these findings matters far more than the count. All 1333 findings are medium-severity IoC matches that are known XIOC garbage patterns. If this were genuinely malicious, we would expect to see malware signatures, obfuscation, or specific suspicious domains—not thousands of property access chains misread as domains. The scoring system has known biases that inflate scores based on IoC volume, which is why the findings count should not drive the verdict. Without actual malware signatures or suspicious behavior, the anonymous status alone does not justify a confirmed_malicious or confirmed_risk verdict.

Key Reasons

• 1329 IoC findings are XIOC extractor garbage (property access chains misread as domains)
• Zero malware signatures detected
• Zero obfuscation findings
• Zero code-smell findings
• Network findings limited to standard Next.js bundled file fetch calls

False Positive Considerations

• XIOC property access chain misinterpretation (fm.finally.call, i.open, self.location.search)
• High finding count driven by IoC volume rather than malicious behavior
• Next.js bundled code triggering standard network findings

Reviewed 2026-05-31; recommended action: suppress false positive; model confidence 75%.