trends.barnas.net

phenomenon · first seen 4 Oct 2026

adversarial attacks on LLM code review

Also: prompt injection on AI code reviewers

Manipulating AI reviewers of pull requests to approve vulnerable code through crafted changes or feedback.

Mentions per week

0119 May11 Aug10 NovPapersCodeDiscussion

PapersCodeDiscussion

Earliest sources

Latest sources

Mentioned together with

vulnerability detection