Most articles about prompt injection quote the same two numbers: OWASP ranks it first in the LLM Top 10, and "most models are vulnerable". Neither tells you how often it actually turns into a registered vulnerability. So we counted.
Since January 2026 the National Vulnerability Database lists 52 CVEs in LLM and agent software where prompt injection is the cause or the consequence (numbers as of 6 November 2026; the full list with NVD links is at promptinjections.de/cve-verzeichnis ).
What the 52 look like
Severity: 6 critical (CVSS 9.0+), 32 high, 13 medium, 1 low. Median base score 8.0.
Timeline: 1 in January, 2 in May, 3 in July, 10 in August, 13 in September, 23 in October. Part of the autumn jump is method: since September we also query the NVD monthly for "prompt injection", so earlier months are under-counted rather than quieter.
Product types: coding agents lead with 8, followed by AI gateways/proxies (6) and MCP servers (6).…