www.glow.io AI Coding Agents Install Vulnerable Deprecated MCP Servers
Article Content
- •AI coding agents frequently install deprecated MCP servers with unpatched vulnerabilities.
- •98% of installations are based on outdated training data, bypassing current secure options.
- •Over 17.3 million downloads of deprecated packages highlight a significant security risk.
Research from Glow Labs reveals that AI coding agents are installing deprecated Model Context Protocol (MCP) servers, which contain high-severity vulnerabilities and will not receive security updates. In 98% of monitored sessions, agents selected these outdated servers based on their training data rather than the latest versions. Notably, one deprecated server has a remote code execution flaw that could allow an attacker to execute arbitrary commands. Despite the risks, these deprecated packages have been downloaded over 17.3 million times since July 2025, indicating widespread use. The vulnerabilities are not tracked by CVEs, making them difficult to detect with standard security measures. Current efforts to warn users about these deprecated packages are inadequate, as agents may ignore deprecation notices. No has been reported yet, but the potential for future attacks remains significant.
Ask AI about this cluster
Answers cite the sources they use
Timeline
More articles in this cluster (3)
Following this threat?
Track Anthropic in your own feed — you're alerted when they show up in new reporting, leak sites or exploitation.
Free account · no card needed
Common questions
What vulnerabilities are associated with the deprecated MCP servers?
Are these vulnerabilities currently being exploited?
What should organizations do to mitigate these risks?
Continue Reading
Critical Zero-Day Exploited in BeyondTrust Remote Support by China-linked Group A critical vulnerability in BeyondTrust Remote Support, tracked as CVE-2026-102764 with a CVSS score of 9.3, has been exploited by the China-linked espionage group Armory Castle since August 29, 2026. This flaw allows unauthenticated remote attackers to perform file operations, gaining root control and access to…
Critical RCE Vulnerability in Prefab UI Disclosed A critical remote code execution vulnerability, CVE-2026-120338, was disclosed on November 1, 2026, affecting Prefab UI's Generative UI sandbox. The flaw allows attackers to execute arbitrary Python code by exploiting Deno's permissions, enabling file read/write access on the host server. This vulnerability affects…