汉字News

← 汉字NewsChinese News for Learners
technology
kē jì
科技
·HSK 6

Cisco
Sīkē
思科
at
zài
在
RSA 2026 (cybersecurity conference)
RSA 2026
RSA 2026
conference
dàhuì
大会
to launch / to release
tuīchū
推出
targeting / aimed at
zhēnduì
针对
artificial intelligence
réngōng zhìnéng
人工智能
agent
dàilǐ
代理
attributive particle
de
的
security framework
ānquán kuàngjià
安全框架
DefenseClaw (security framework)
DefenseClaw
DefenseClaw
。

at
zài
在
RSA 2026 (cybersecurity conference)
RSA 2026
RSA 2026
at the conference
dàhuì shàng
大会上
，
Cisco
Sīkē
思科
to release / to launch
fābù
发布
completed action particle
le
了
targeting
zhēnduì
针对
artificial intelligence
réngōng zhìnéng
人工智能
agent
dàilǐ
代理
attributive particle
de
的
security framework
ānquán kuàngjià
安全框架
DefenseClaw
DefenseClaw
DefenseClaw
，
to integrate
zhěnghé
整合
multiple / various
duō zhǒng
多种
open-source
kāiyuán
开源
security tools
ānquán gōngjù
安全工具
。
furthermore / in addition
cǐwài
此外
，
Cisco
Sīkē
思科
also will
hái jiāng
还将
zero trust
líng xìnrèn
零信任
access
fǎngwèn
访问
mechanism
jīzhì
机制
to extend to
kuòzhǎn zhì
扩展至
artificial intelligence
réngōng zhìnéng
人工智能
agents
dàilǐ
代理
，
and launched
bìng tuīchū
并推出
AI Defense Explorer Edition
AI Defense Tànsuǒ Bǎn
AI Defense探索版
for
wèi
为
developers
kāifāzhě
开发者
to provide
tígōng
提供
security testing tools
ānquán cèshì gōngjù
安全测试工具
。

March 25, 2026 at 8:00 PM

at / in
zài
在
RSA 2026 (cybersecurity conference)
RSA 2026
RSA 2026
cybersecurity
wǎngluò ānquán
网络安全
conference
dàhuì
大会
at (location particle)
shàng
上
，
Cisco
Sīkē
思科
to release / to launch
fābù
发布
completed action particle
le
了
specifically for
zhuān wèi
专为
artificial intelligence
réngōng zhìnéng
人工智能
agent
dàilǐ
代理
to design / designed
shèjì
设计
attributive particle
de
的
security framework
ānquán kuàngjià
安全框架
DefenseClaw (security framework)
DefenseClaw
DefenseClaw
。
this / the said
gāi
该
framework
kuàngjià
框架
to integrate / to consolidate
zhěnghé
整合
completed action particle
le
了
Skills Scanner (tool)
jìnéng sǎomiáo qì
技能扫描器
、
MCP Scanner (tool)
MCP sǎomiáo qì
MCP扫描器
、
AI asset inventory
réngōng zhìnéng zīchǎn qīngdān
人工智能资产清单
and / as well as
jí
及
code protection
dàimǎ fánghù
代码防护
and so on / etc.
děng
等
open-source tools
kāiyuán gōngjù
开源工具
，
to ensure
quèbǎo
确保
each / every
měi gè
每个
agent
dàilǐ
代理
before
zài
在
deployment
bùshǔ
部署
before
qián
前
to undergo
jīngguò
经过
review / inspection
shěnchá
审查
and
yǔ
与
sandbox isolation
shābèi gélí
沙盒隔离
。
furthermore / in addition
cǐwài
此外
，
Cisco
Sīkē
思科
to (formal preposition)
jiāng
将
zero trust
líng xìnrèn
零信任
access mechanism
fǎngwèn jīzhì
访问机制
to extend / to expand
kuòzhǎn
扩展
to / towards
zhì
至
artificial intelligence
réngōng zhìnéng
人工智能
agents
dàilǐ
代理
，
to require
yāoqiú
要求
each
měi gè
每个
agent
dàilǐ
代理
with
yǔ
与
specific / designated
tèdìng
特定
human
rénlèi
人类
employee
yuángōng
员工
account
zhànghù
账户
to bind / to link
bǎngdìng
绑定
，
and / also
bìng
并
through / via
tōngguò
通过
MCP (Model Context Protocol)
MCP
MCP
gateway
wǎngguān
网关
to route
lùyóu
路由
all
suǒyǒu
所有
traffic / data flow
liúliàng
流量
。
Cisco
Sīkē
思科
simultaneously / at the same time
tóngshí
同时
to launch
tuīchū
推出
completed action particle
le
了
AI Defense Explorer Edition
AI Defense Tànsuǒ Bǎn
AI Defense探索版
，
for
wèi
为
developers
kāifāzhě
开发者
to provide
tígōng
提供
self-service
zìzhù
自助
testing tools
cèshì gōngjù
测试工具
，
in order to
yǐ
以
to evaluate / to assess
pínggū
评估
model
móxíng
模型
when
zài
在
under attack
zāoshòu gōngjī
遭受攻击
when / at the time of
shí
时
attributive particle
de
的
resilience / robustness
rènxìng
韧性
。
Cisco
Sīkē
思科
survey
diàochá
调查
to show / to indicate
xiǎnshì
显示
，
85 percent
bǎifēnzhī bāshíwǔ
百分之八十五
attributive particle
de
的
large-scale
dàxíng
大型
enterprise / company
qǐyè
企业
are already
yǐ zài
已在
experimenting with
shìyàn
试验
AI agents
réngōng zhìnéng dàilǐ
人工智能代理
，
however / yet
rán ér
然而
only
jǐn yǒu
仅有
five percent
bǎifēnzhī wǔ
百分之五
have
jiāng
将
them / it
qí
其
to deploy / to put into use
tóurù
投入
actual / real
shíjì
实际
production environment
shēngchǎn huánjìng
生产环境
。
analysts
fēnxī rénshì
分析人士
to point out
zhǐchū
指出
，
as / along with
suízhe
随着
enterprises
qǐyè
企业
to accelerate
jiāsù
加速
to deploy
bùshǔ
部署
AI agents
réngōng zhìnéng dàilǐ
人工智能代理
，
security protection system
ānquán fánghù tǐxì
安全防护体系
attributive particle
de
的
lag / lagging behind
zhìhòu
滞后
already
yǐ
已
to become
chéngwéi
成为
industry
hángyè
行业
adoption / promotion
tuīguǎng
推广
attributive particle
de
的
primary obstacle
zhǔyào zhàng ài
主要障碍
。
given that / in light of
jiànyú
鉴于
AI agents
réngōng zhìnéng dàilǐ
人工智能代理
to be able to / can
nénggòu
能够
on behalf of
dàibiǎo
代表
user
yònghù
用户
to autonomously execute
zìzhǔ zhíxíng
自主执行
complex
fùzá
复杂
tasks
rènwu
任务
，
once / if ever
yīdàn
一旦
passive marker
bèi
被
attacker
gōngjī zhě
攻击者
to hijack
jiéchí
劫持
，
consequences
hòuguǒ
后果
will
jiāng
将
to far exceed
yuǎn chāo
远超
traditional
chuántǒng
传统
software vulnerability
ruǎnjiàn lòudòng
软件漏洞
。

New Words

Sīkē
思科
Cisco
ānquán kuàngjià
安全框架
security framework
dàilǐ
代理
agent
shābèi gélí
沙盒隔离
sandbox isolation
líng xìnrèn fǎngwèn jīzhì
零信任访问机制
zero trust access
bùshǔ
部署
deployment
shēngchǎn huánjìng
生产环境
production environment
ruǎnjiàn lòudòng
软件漏洞
software vulnerability

News in English

Science & Technology

Cisco Unveils DefenseClaw Security Framework for AI Agents at RSA Conference 2026

Cisco launched DefenseClaw at RSA Conference 2026, a security framework for AI agents that integrates open-source tools and extends Zero Trust access controls to AI agents for the first time.

At the RSA 2026 cybersecurity conference, Cisco launched DefenseClaw, a security framework designed for AI agents. The framework integrates the Skills Scanner, MCP Scanner, AI asset inventory, and code protection tools, ensuring every agent undergoes review and sandbox isolation before deployment.

Cisco also extends zero trust access mechanisms to AI agents, requiring each agent to bind to specific human employee accounts and routing all traffic through an MCP gateway.

Simultaneously, Cisco launched the AI Defense Explorer Edition, providing developers self-service testing tools to evaluate model resilience under attack.

Cisco's survey shows that 85 percent of large enterprises are experimenting with AI agents, yet only 5 percent have deployed them in actual production environments.

Analysts point out that as enterprises accelerate AI agent deployment, the lagging security protection system has become the industry's primary adoption obstacle.

Given that AI agents can autonomously execute complex tasks on behalf of users, the consequences of hijacking would far exceed those of traditional software vulnerabilities.

More in Science & Technology

China
Zhongguó
中国
AI
réngong zhìnéng
人工智能
models
móxíng
模型
captures
zhànlǐng
占领
market
shìchǎng
市场

Chinese AI Models Gain Market Share as GLM 5.2 Rivals OpenAI and Anthropic at Low Cost

UBTech
Yōubìxuǎn
优必选
U1
U1
U1
humanoid robot
rénxíng jīqìrén
人形机器人
launches
fābuō
发布
price
jiàgé
价格

UBTech Launches U1 Ultra-Bionic Humanoid Robot Series Starting at 17600 Dollars

Apple (company)
Píngguǒ
苹果
to file a lawsuit / to sue
qǐsù
起诉
OpenAI (AI company)
OpenAI
OpenAI
，
to accuse / to charge
zhǐkòng
指控
it / the company (formal)
qí
其
to steal / to misappropriate
qièqǔ
窃取
hardware
yìngjiàn
硬件
trade secrets / confidential business information
shāngyè jīmì
商业机密
。

Apple Sues OpenAI Alleging Hardware Trade Secret Theft

Tencent (tech company)
Téngxùn
腾讯
in
zài
在
WeChat (messaging app)
Wēixìn
微信
to launch / roll out
tuīchū
推出
artificial intelligence
réngōng zhìnéng
人工智能
assistant
zhùshǒu
助手
Xiaowei (AI assistant name)
Xiǎowēi
小微
。

Tencent Begins Rolling Out AI Agent Xiaowei to WeChat Users