Context Privilege Escalation paper

Evidence record

Read the original source ↗

Claim-level citations (4)

• supportsContext privilege escalation let repository instructions blind an AI code reviewer: The CPE researchers report that a malicious AGENTS.override.md caused Codex review to approve manipulated vulnerable code, while a control review without the override detected the command injection.

  V-D Manipulated Pull-Request Review in Codex, attack and control comparison

• supportsContext privilege escalation let repository instructions blind an AI code reviewer: Researchers evaluated 12 real agent harnesses and reported consequences including manipulated tool invocation, denial of service, remote code execution and full agent compromise.

  Abstract, scope of twelve agent harnesses

• supportsContext privilege escalation let repository instructions blind an AI code reviewer: In the Codex demonstration, attacker-controlled repository instructions caused Codex to approve vulnerable code that it rejected in a control run without the malicious override.

  V-D Manipulated Pull-Request Review in Codex, attack and control comparison

• supportsContext privilege escalation let repository instructions blind an AI code reviewer: The paper defines MessageRole Context Privilege Escalation and Cross-Scope Context Privilege Escalation as two classes of context-assembly attack.

  Abstract; III Context Privilege Escalations in LLM Agents