>
正文
Hermes Agent reshapes the security paradigm of AI agents within 42 days through its self-evolving architecture
2026-04-09 18:40

On February 25, 2026, Nous Research released Hermes Agent v0.1.0. Forty-two days later, on April 8, the project had been updated to v0.8.0, involving a total of 8 major version releases, hundreds of pull requests, and contributions from 242 contributors.

Meanwhile, the OpenClaw project on GitHub, which had received 346,000 stars, exposed 138 security vulnerabilities within 63 days, of which 41% were classified as severe or high-risk, including a zero-click remote code execution vulnerability with a CVSS score of 8.8, CVE-2026-25253.

According to Shodan scan data, in February, over 42,000 OpenClaw instances were exposed to the public internet, and 63% of them did not have gateway authentication enabled, highlighting the critical flaws in the default security configurations of current mainstream proxy architectures.

免责声明:本内容为作者独立观点,不代表平台立场。未经允许不得转载,文中内容仅供参考,不作为实际操作建议,交易风险自担。
标签:
Claude Code
Solayer Labs
Fuzzland
React
Kubernetes
Honcho
Eden Network
Hermes 4.3
Psyche
agentskills.io
ClawHub
Shodan
Blink Security Blog
FTS5
Hermes 1
Hermes Agent
OpenClaw
Snyk
Hermes 4
Anthropic
OpenAI
OSS Capital
VentureBeat
Distributed Global
Nous R
分享: