Where AI Vulnerability Discovery Harnesses Go Quiet: Structural False Negatives and Hidden Vulnerabilities

Reported by arXiv cs.CR (Gadi Evron)

Summary by Cyber Security News22 Oct 2026, 06:00 CET

About this summary. This is a short excerpt from the description that arXiv cs.CR published with its article. Cyber Security News did not report or verify the underlying story. Read the original: https://arxiv.org/abs/2610.29379

arXiv:2610.29379v1 Announce Type: new Abstract: LLM-based vulnerability discovery harnesses can miss real, reachable vulnerabilities because of how the harness around the model is built. Steps that select code, filter it, and store and merge findings can drop a vulnerability before or after the model runs, independently of the model's capability, and the report usually gives no sign that this happened. These drops come from engineering choices, and errors.

Read the full article at arXiv cs.CR →

Source attribution: headline and facts are from arXiv cs.CR (arxiv.org). Summary method: excerpt of the source description. See our source attribution policy.

Related coverage

Threat Actor Abuses Agentic AI Pentesting Framework to Target South Korean Financial Institutions

Log-To-Leak: Prompt Injection Attacks on Tool-Using LLM Agents via the Model Context Protocol

SpecGuard: Detecting Multi-Turn LLM Jailbreaks via Specification-Based Runtime Monitoring

LLM agent security needs more than guardrails

A Deterministic Security Runtime for LLM Agents on MPI and Batch-Scheduled HPC Systems: Consensus-Governed, Hardware-Rooted Supervision of Every Rank

Secure and Functional Code Generation with Self-Evaluation