Unverified50% confidenceCautionExact time
Security researchers say Claude Code's default auto mode can be hijacked by prompt injection in web summary requests, thereby triggering code execution, with small-sample tests showing an attack success rate as high as 60% to 80%.
1
Sources
50%
Confidence
Medium-term (~90 days)
Relevance
9/16/2026
First Seen
Valid until: 12/15/2026
Sources
Related Entities
Related Claims
UnverifiedIn adversarial testing, synthetic attacks still have approximately a 7% miss rate against Claude Code's safety classifier.72% similarUnverifiedClaude Code将自动模式设为Pro/Max/Team用户默认权限模式，用独立分类器审查Shell命令，测试中拦截了89%的危险命令，而手动审批仅拦截14%70% similarUnverifiedClaude Code等产品表现出较强拒绝能力部分源于训练阶段引入大量对抗样本进行安全对齐，基于规则过滤的Agent在对抗性测试中更容易被绕过67% similarUnverifiedTraditional code review processes can capture approximately 60% of code defects.66% similarUnverifiedAnthropic has provided a misjudgment feedback mechanism for Claude Code, allowing users to submit appeal reports through a built-in feedback channel when a request is incorrectly flagged, in order to calibrate the safety classifier.66% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/927560API
curl https://kongchang.com/api/v1/knowledge/claims/927560MCP
get_claim(id=927560)