GoKawiilOWASP researchers find prompt injection risk invisible to security scanners
OWASP GenAI Security Project leaders Kyriakos Lambros and Steve Wilson published an analysis comparing expert rankings of LLM security threats against 6,639 real-world incidents pulled from CVE, GitHub advisories, OSV and the AIAAIC database. Prompt injection tops the OWASP expert list for three straight years but ranks only 12th in the incident data, and statistical testing found no reliable agreement between the two rankings.
GoKawiil's interpretation of the reporting above, not reported fact.
The mismatch suggests that standard vulnerability scanners and public incident databases systematically undercount prompt injection risk because the attack exploits model behavior rather than a traceable code flaw. CISOs relying on CVE counts to prioritize AI security spending could be dangerously underestimating the top-rated threat in the industry's own expert consensus.
- Prompt injection ranks No. 1 among security experts but only No. 12 in logged real-world incidents.
- Statistical analysis (Cohen's kappa of 0.20) found no reliable agreement between expert opinion and incident data.
- The gap exists because prompt injection attacks leave no trace detectable by conventional vulnerability scanners.
The Developer's Playbook for Large Language Model Security by Steve Wilson — Written by Steve Wilson, one of the OWASP Top 10 for LLM Applications leaders cited in this analysis, this O'Reilly book digs into prompt injection and the other risks that scanners simply can't see. It's a practical way for security teams to understand why the CVE scoreboard underrates LLM-specific attacks.
See The Developer's Playbook for Large Language Model Security by Steve Wilson on Amazon → Affiliate link — we may earn a commission on purchases, at no extra cost to you. Product picked by AI based on this article; it is not a tested recommendation.Source: venturebeat.com, 2026-08-25
Published there as: “Prompt injection ranks No. 1 with OWASP and No. 12 in the incident record. The attack itself is invisible to a scan.”
Read the original report → The summary and analysis above are GoKawiil's own, written from reporting by the source above. Facts and quotes belong to the original publisher.