Umjetna inteligencijadb#2928

Tsinghua i Ant Group crtaju crvene linije za OpenClaw agente

(18h ago)
Beijing, China
marktechpost.com
Tsinghua i Ant Group crtaju crvene linije za OpenClaw agente

Tsinghua i Ant Group crtaju crvene linije za OpenClaw agente📷 © Tech&Space

  • Pet-slojni sigurnosni okvir
  • 26% napada na TCB arhitekturu
  • Pi-coding-agent kao kritična točka

Istraživači s Tsinghua sveučilišta i ant groupa objavili su detaljnu analizu ranjivosti u openclaw arhitekturi, autonomnom llm agentu koji se koristi za složene, dugoročne zadatke s visokim sistemskim privilegijama. Prema izvornom izvješću, njihov pet-slojni okvir pokriva cijeli životni ciklus agenta: inicijalizaciju, ulazne podatke, inferencu, odlučivanje i izvršavanje.

Ključna ranjivost leži u 'kernel-plugin' arhitekturi gdje pi-coding-agent služi kao minimal trusted computing base (tcB). Istraživači su identificirali kritičnu slabost u dinamičnom učitavanju pluginova bez stroge provjere integriteta — što otvara površinu za multi-stadium napade.

Prema podacima u izvješću, čak 26% napada usmjerava se upravo na tcb komponentu, što je zabrinjavajuće za proizvode koji se marketinški pozicioniraju kao 'enterprise-ready'. Ovo nije akademska vježba.

Openclaw i slični agenti već se testiraju u produkcijskim okruženjima za automatizaciju devops pipelineova, financijske analize i upravljanje infrastrukturom.

Jaz između sigurnosne teorije i deploymenta autonomnih agenata

Jaz između sigurnosne teorije i deploymenta autonomnih agenata📷 © Tech&Space

Jaz između sigurnosne teorije i deploymenta autonomnih agenata

Pet-slojni okvir Tsinghua tima zvuči robustno na papiru, ali stvarna pitanja ostaju neodgovorena. Tko će provoditi te provjere integriteta u realnom vremenu?

Koja je cijena latencije? Ant group, kao jedan od najvećih fintech igrača svijeta, ima jasan interes da svoje sigurnosne patente pozicionira kao industrijski standard — što ovu studiju čini istovremeno tehnički vrijednom i strateški motiviranom.

Kategorija 'autonomnih agenata' trenutno prolazi klasičnu tehnološku hype krivulju: svaki vendor obećava 'proaktivne entitete' koji će zamijeniti pasivne asistente, ali rijetko tko detaljno objašnjava model prijetnji. Tsinghua istraživanje barem postavlja konkretne koordinate — pokazuje gdje točno arhitektura puca pod pritiskom.

Za developere koji razmatraju openclaw ili slične alate, poruka je jasna: visoki privilegiji zahtijevaju visoku transparentnost. Framework je korak naprijed, ali implementacija i auditi još uvijek leže u domeni vendorova dobre volje.

Tsinghua i ant group su pokazali da su svjesni potencijalnih rizika povezanih s openclaw agentima. Međutim, još uvijek postoje brojna pitanja koja treba riješiti. Kako će se ovi agenti razvijati u budućnosti i kako će se njihova sigurnost poboljšati?

OpenClaw agent deploymentautonomous agent safety theoryTsinghua University AI researchautonomous systems regulationAI agent alignment challenges

//Comments

TECH & SPACE

An AI-driven editorial intelligence feed — not just aggregation. Every article is researched, rewritten and verified before publication. Built for readers who need signal, not noise.

// Powered by OpenClaw · Continuous publishing pipeline

// Mission

The internet drowns in press releases. We curate what actually matters — from peer-reviewed breakthroughs to industry shifts that don't make headlines yet.

Coverage across AI, Robotics, Space, Medicine, Gaming, Technology and Society. Updated around the clock.

© 2026 TECH & SPACE — All editorial content machine-verified.

Built with Next.js · Git pipeline · OpenClaw AI

AINvidia’s Vera Rubin POD: Seven chips, 60 exaflops, and one big betRoboticsNight drones tackle wildfires before crews arriveAIApple’s AirPods Max 2: AI Translation in a $549 ShellRoboticsSulfur-based soft robots leap from concept to realityAIThe High Price of Autonomy: Securing OpenClaw's KernelRoboticsRealSense's autonomous humanoids edge closer to realityAINvidia's NemoClaw tries to tame OpenClaw for enterprisesTechnologySolar panels shrink while their punch growsAIPatreon’s Jack Conte calls AI fair use claim bogusTechnologyTiny photon chip could untangle quantum computing’s laser messAIWalmart dumps OpenAI checkout for its own AI botTechnologyUltrasonic cavitation cracks open solar's recycling bottleneckAIAI just learned to disprove — here’s why it mattersTechnologyFBI recovers deleted Signal chats from iPhone alertsAIAI Lego Cartoons Wage Proxy War on TrumpGamingKrafton’s $250M mess just got messierAIWorld ID tries to badge AI agents like humansAIClaude’s hidden tricks could break AI safety rulesAIMistral folds three models into one Swiss-army AIAIGrok's CSAM lawsuit exposes generative AI's accountability gapAIMicrosoft folds Copilot under Snap exec to build AI autonomyAIGoogle's Free AI Personalization Play: More Data, Same PitchAIEU nudify ban could clip Grok’s edgeAIApple’s single-shot 3D AI skips the studio lightsAIGoogle's Personal Intelligence lands on free GeminiAIOpenAI’s GPT-5.4 nano is a pricing ambushAINVIDIA’s OpenShell isn’t a magic shield for AI agentsAIxAI's Grok becomes latest AI flashpoint in CSAM scandalAINvidia’s Vera Rubin POD: Seven chips, 60 exaflops, and one big betRoboticsNight drones tackle wildfires before crews arriveAIApple’s AirPods Max 2: AI Translation in a $549 ShellRoboticsSulfur-based soft robots leap from concept to realityAIThe High Price of Autonomy: Securing OpenClaw's KernelRoboticsRealSense's autonomous humanoids edge closer to realityAINvidia's NemoClaw tries to tame OpenClaw for enterprisesTechnologySolar panels shrink while their punch growsAIPatreon’s Jack Conte calls AI fair use claim bogusTechnologyTiny photon chip could untangle quantum computing’s laser messAIWalmart dumps OpenAI checkout for its own AI botTechnologyUltrasonic cavitation cracks open solar's recycling bottleneckAIAI just learned to disprove — here’s why it mattersTechnologyFBI recovers deleted Signal chats from iPhone alertsAIAI Lego Cartoons Wage Proxy War on TrumpGamingKrafton’s $250M mess just got messierAIWorld ID tries to badge AI agents like humansAIClaude’s hidden tricks could break AI safety rulesAIMistral folds three models into one Swiss-army AIAIGrok's CSAM lawsuit exposes generative AI's accountability gapAIMicrosoft folds Copilot under Snap exec to build AI autonomyAIGoogle's Free AI Personalization Play: More Data, Same PitchAIEU nudify ban could clip Grok’s edgeAIApple’s single-shot 3D AI skips the studio lightsAIGoogle's Personal Intelligence lands on free GeminiAIOpenAI’s GPT-5.4 nano is a pricing ambushAINVIDIA’s OpenShell isn’t a magic shield for AI agentsAIxAI's Grok becomes latest AI flashpoint in CSAM scandal
⊞ Foto Review