AI Spera launches AITEM security tool to find exposed servers and conduct web penetration testing
November 20, 2026 · 3 min read
Also published in עברית
AI Spera released AITEM on November 19, a security product that combines artificial intelligence agents, public vulnerability impact analysis, and web penetration testing with its existing attack surface management capabilities. The tool helps organizations identify internet-exposed servers and websites, determine which vulnerabilities affect them, identify who manages them, and prioritize inspection points.
The product addresses a common security challenge: servers exposed to the internet that do not appear in asset inventories. During a press briefing on November 18, AI Spera chief executive Kang Byung-tak demonstrated how the system identified an exposed server and inferred its operator with 63 percent confidence by searching internal wikis and work tickets for access alert handling records. Security teams can review the reasoning behind the inference. Kang noted that security staff often spend days searching multiple systems with the same IP address to identify a single server's operator.
AITEM includes a CVE Analyzer feature that cross-references publicly disclosed vulnerabilities mentioned in domestic and international security news with a company's externally exposed assets. The tool distinguishes whether a company uses a product mentioned in news reports and whether the corresponding vulnerability was detected. News items are categorized into four levels based on relevance to the company's assets. The system also incorporates Criminal IP threat intelligence to examine external exposure status, malicious infrastructure, and attack activity information.
The product supports natural language requests for repetitive tasks. Users can request actions such as sending information about externally exposed database servers to a messaging platform daily at midnight, and the agent configures query conditions and notification schedules. Integration with company firewalls and security information and event management systems enables detection of unauthorized artificial intelligence use by employees outside organizational approval boundaries.
AITEM's management scope includes external exposure of assets belonging to subsidiaries, partners, and suppliers. The system detects sensitive files, application programming interface keys, and authentication credentials exposed in search engines and public code repositories. It also identifies compromised accounts and breach indicators related to the company on dark web platforms for further investigation.
The tool performs web penetration testing on discovered websites included in management scope. Kang described an example involving a customer consultation inquiry web service where missing access permission checks allow users to view other people's consultation records. Such defects in company-developed web services are difficult to identify by checking public vulnerability lists alone. As generative artificial intelligence for code writing spreads, security teams must examine development and testing sites previously unknown to them. Kang stated that one large enterprise had over 100 websites discovered during external asset exploration. The company structured inspection items around frequently occurring defects such as access permission validation omissions, targeting completion in under 30 minutes per site. The system applies non-destructive testing methods that exclude checks that could alter data or create service load.
AI Spera shared its own response to a critical Node.js vulnerability as a case study. The company initially identified several hundred exposed servers through its existing attack surface management system but discovered additional servers managed by departments outside the security organization during response efforts. The company applied patches or blocked access through firewalls within one day for internet-exposed servers. Some internal servers required replacement due to ended operating system support, and decisions remained regarding service interruption timing and server restart scheduling. Kang explained that AITEM can accelerate initial inspection of externally accessible assets, though separate work is necessary for internal asset identification and patch scheduling coordination.
The company projected revenue growth for the current year compared to the previous year. After noting that last year's revenue doubled compared to the prior year, Kang stated that this year's growth rate would not reach double. He did not disclose specific revenue projections. Kang noted that while security industry revenue concentrates in November and December, the company has increased its overseas customer base this year. AI Spera introduced AITEM at security conferences in Britain and Singapore before its domestic official launch.