Unverified50% confidenceBenchmarkExact time
2023年安全研究机构Vulcan Cyber的研究报告显示，主流AI编程助手中幻觉包名的出现率在特定场景下高达21.3%
1
Sources
50%
Confidence
Long-term
Relevance
7/10/2026
First Seen
Sources
AI Coding Security: A CLI Tool Designed to Block Vulnerable Dependencies for Agents
hackernewshackernews7/1/2026
Related Claims
UnverifiedAcademic research indicates that current state-of-the-art AI detectors can have false negative rates as high as 40%–60% when faced with AI-generated text that has been simply paraphrased.73% similarUnverifiedCompanies including Meta, Google, and Microsoft have successively disclosed internal AI coding penetration rate data, with some teams having more than 30% of their code generated by AI.73% similarUnverified部分计算机科学领域的会议报告称超过15%的投稿存在明显的AI生成痕迹72% similarUnverifiedResearchers analyzed more than 42,000 AI Agent Skills, of which 26.1% had security vulnerabilities and 5.2% showed suspected malicious signs.72% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/412472API
curl https://kongchang.com/api/v1/knowledge/claims/412472MCP
get_claim(id=412472)