New approach to agent reliability, AgentSpec, forces agents to follow rules
Researchers from Singapore Management University developed a new domain-specific language for agents to remain reliable.
#ai #airesearch #news
Researchers from Singapore Management University developed a new domain-specific language for agents to remain reliable.
#ai #airesearch #news
New approach to agent reliability, AgentSpec, forces agents to follow rules
Researchers from Singapore Management University developed a new domain-specific language for agents to remain reliable.
venturebeat.com
March 29, 2025 at 10:12 PM
New approach to agent reliability, AgentSpec, forces agents to follow rules
Researchers from Singapore Management University developed a new domain-specific language for agents to remain reliable.
#ai #airesearch #news
Researchers from Singapore Management University developed a new domain-specific language for agents to remain reliable.
#ai #airesearch #news
Building custom UIs for AI agents usually means rebuilding the same "middle layer" in every stack to keep state and progress in sync. We are changing that 🧵
March 16, 2026 at 4:56 PM
Building custom UIs for AI agents usually means rebuilding the same "middle layer" in every stack to keep state and progress in sync. We are changing that 🧵
Started using Semble to save research on AI agent governance.
My collection:
https://semble.so/profile/lasa.numina.systems
Papers on AgentSpec, MI9, Runtime Governance. Data lives on my PDS — AT Proto is neat.
My collection:
https://semble.so/profile/lasa.numina.systems
Papers on AgentSpec, MI9, Runtime Governance. Data lives on my PDS — AT Proto is neat.
April 15, 2026 at 3:48 AM
Started using Semble to save research on AI agent governance.
My collection:
https://semble.so/profile/lasa.numina.systems
Papers on AgentSpec, MI9, Runtime Governance. Data lives on my PDS — AT Proto is neat.
My collection:
https://semble.so/profile/lasa.numina.systems
Papers on AgentSpec, MI9, Runtime Governance. Data lives on my PDS — AT Proto is neat.
venturebeat.com/ai/new-appro...
@enilev.bsky.social
my enilev🪔mom
Mr @tonymoroney.bsky.social
🌐 #BOSCh
@bsky.app #blueSky
#bigdata #ML #usa
#CyberSecurity #Technology #uk #bitcoin #genAI #ai #eUrope #frenchTECh #Linkedin #HealthTech #DiGiTALhealth #5g #Quantum #Health #web3 #SOciaLmeDia
@enilev.bsky.social
my enilev🪔mom
Mr @tonymoroney.bsky.social
🌐 #BOSCh
@bsky.app #blueSky
#bigdata #ML #usa
#CyberSecurity #Technology #uk #bitcoin #genAI #ai #eUrope #frenchTECh #Linkedin #HealthTech #DiGiTALhealth #5g #Quantum #Health #web3 #SOciaLmeDia
New approach to agent reliability, AgentSpec, forces agents to follow rules
Researchers from Singapore Management University developed a new domain-specific language for agents to remain reliable.
venturebeat.com
April 16, 2025 at 5:44 PM
Chen, Shen, Kang, Hong, Jiang, Bose, Zhang, Leng, Vyas, Mao, Ouyang, Zhou, Qin: AgentSpec: Understanding Embodied Agent Scaffolds Through Controlled Composition https://arxiv.org/abs/2606.14674 https://arxiv.org/pdf/2606.14674 https://arxiv.org/html/2606.14674
June 15, 2026 at 6:40 AM
Chen, Shen, Kang, Hong, Jiang, Bose, Zhang, Leng, Vyas, Mao, Ouyang, Zhou, Qin: AgentSpec: Understanding Embodied Agent Scaffolds Through Controlled Composition https://arxiv.org/abs/2606.14674 https://arxiv.org/pdf/2606.14674 https://arxiv.org/html/2606.14674
Jixuan Chen, Jianzhi Shen, Haoqiang Kang, Zhi Hong, Qingyi Jiang, Soham Bose, Yiming Zhang, Leon Leng, Amit Vyas, Lingjun Mao, Siru Ouyang, Kun Zhou, Lianhui Qin
AgentSpec: Understanding Embodied Agent Scaffolds Through Controlled Composition
https://arxiv.org/abs/2606.14674
AgentSpec: Understanding Embodied Agent Scaffolds Through Controlled Composition
https://arxiv.org/abs/2606.14674
June 15, 2026 at 12:56 PM
Jixuan Chen, Jianzhi Shen, Haoqiang Kang, Zhi Hong, Qingyi Jiang, Soham Bose, Yiming Zhang, Leon Leng, Amit Vyas, Lingjun Mao, Siru Ouyang, Kun Zhou, Lianhui Qin
AgentSpec: Understanding Embodied Agent Scaffolds Through Controlled Composition
https://arxiv.org/abs/2606.14674
AgentSpec: Understanding Embodied Agent Scaffolds Through Controlled Composition
https://arxiv.org/abs/2606.14674
🛡️ AgentToolApprovalsExample: human-in-the-loop, in 4 lines.
id: demo-tool-approvals
tools:
- runtime-echo:0.0.1 # auto
- runtime-sensitive-echo:0.0.1 # manual approval
One agentspec, two tools — runtime-echo ships, runtime-sensitive-echo waits for your 👍.
#AIAgents #Datalayer
id: demo-tool-approvals
tools:
- runtime-echo:0.0.1 # auto
- runtime-sensitive-echo:0.0.1 # manual approval
One agentspec, two tools — runtime-echo ships, runtime-sensitive-echo waits for your 👍.
#AIAgents #Datalayer
April 28, 2026 at 4:32 PM
🛡️ AgentToolApprovalsExample: human-in-the-loop, in 4 lines.
id: demo-tool-approvals
tools:
- runtime-echo:0.0.1 # auto
- runtime-sensitive-echo:0.0.1 # manual approval
One agentspec, two tools — runtime-echo ships, runtime-sensitive-echo waits for your 👍.
#AIAgents #Datalayer
id: demo-tool-approvals
tools:
- runtime-echo:0.0.1 # auto
- runtime-sensitive-echo:0.0.1 # manual approval
One agentspec, two tools — runtime-echo ships, runtime-sensitive-echo waits for your 👍.
#AIAgents #Datalayer
Xin Wang, Ziming Miao, Yi Zhu, Hui Shen, Zhongwei Wan, Fan Yang, Mi Zhang: AgentSpec: Speculative Decoding for Batch Inference of LLM Agents https://arxiv.org/abs/2608.24004 https://arxiv.org/pdf/2608.24004 https://arxiv.org/html/2608.24004
August 26, 2026 at 6:39 AM
Xin Wang, Ziming Miao, Yi Zhu, Hui Shen, Zhongwei Wan, Fan Yang, Mi Zhang: AgentSpec: Speculative Decoding for Batch Inference of LLM Agents https://arxiv.org/abs/2608.24004 https://arxiv.org/pdf/2608.24004 https://arxiv.org/html/2608.24004
Новый подход к надежности агентов, AgentSpec, заставляет агентов следовать правилам
Исследователи из Университета менеджмента Сингапура разработали новый язык для областей применения, чтобы агенты оставались надежными.
#ai #news
Исследователи из Университета менеджмента Сингапура разработали новый язык для областей применения, чтобы агенты оставались надежными.
#ai #news
New approach to agent reliability, AgentSpec, forces agents to follow rules
venturebeat.com
April 2, 2025 at 1:26 PM
https://venturebeat.com/ai/new-approach-to-agent-reliability-agentspec-forces-agents-to-follow-rules/
#AI #Agentic #AI #agents #agents #sdk #AI #agent #AI #agents #AI
Event Attributes
March 28, 2025 at 11:15 PM
https://venturebeat.com/ai/new-approach-to-agent-reliability-agentspec-forces-agents-to-follow-rules/
#AI #Agentic #AI #agents #agents #sdk #AI #agent #AI #agents #AI
Event Attributes
March 28, 2025 at 10:14 PM
https://venturebeat.com/ai/new-approach-to-agent-reliability-agentspec-forces-agents-to-follow-rules/
#AI #Agentic #AI #agents #agents #sdk #AI #agent #AI #agents #AI
Event Attributes
March 28, 2025 at 9:36 PM
The Open Agent Specification, a framework‑agnostic declarative language for AI agents, was published on 5 Oct 2025 to enable interchangeable workflows across multiple AI frameworks. https://getnews.me/open-agent-specification-boosts-ai-agent-interoperability/ #agentspec #aiinteroperability
October 7, 2025 at 11:16 PM
The Open Agent Specification, a framework‑agnostic declarative language for AI agents, was published on 5 Oct 2025 to enable interchangeable workflows across multiple AI frameworks. https://getnews.me/open-agent-specification-boosts-ai-agent-interoperability/ #agentspec #aiinteroperability
💎AgentSpecは知覚・記憶・推論・行動を型付きで分けて組める。SIMMERは“動いてるのに後で詰む”潜在失敗を測って、frontierモデルでもerror-free最大17%。 https://arxiv.org/abs/2606.14674 https://arxiv.org/abs/2606.14574
June 15, 2026 at 9:38 PM
💎AgentSpecは知覚・記憶・推論・行動を型付きで分けて組める。SIMMERは“動いてるのに後で詰む”潜在失敗を測って、frontierモデルでもerror-free最大17%。 https://arxiv.org/abs/2606.14674 https://arxiv.org/abs/2606.14574
🚀 Say goodbye to rogue AI agents! AgentSpec is a groundbreaking framework that ensures AI safety and reliability in our daily tasks. Curious how it works? Dive in and join the conversation! #AI #AgentSpec #AutomationMagic LINK
March 28, 2025 at 10:18 PM
🚀 Say goodbye to rogue AI agents! AgentSpec is a groundbreaking framework that ensures AI safety and reliability in our daily tasks. Curious how it works? Dive in and join the conversation! #AI #AgentSpec #AutomationMagic LINK
To comply with the rules of new approach to Agent Reliability, Agentspec, Force Agents
Join our daily and weekly newsletters for the latest updates and exclusive content in the industry's leading AI coverage. Learn more AI agents have security and reliability problems. Although the agents allow…
Join our daily and weekly newsletters for the latest updates and exclusive content in the industry's leading AI coverage. Learn more AI agents have security and reliability problems. Although the agents allow…
To comply with the rules of new approach to Agent Reliability, Agentspec, Force Agents
Join our daily and weekly newsletters for the latest updates and exclusive content in the industry's leading AI coverage. Learn more AI agents have security and reliability problems. Although the agents allow enterprises to take more steps in the workflows, performing a task, it is not very flexible and difficult to manage. Organizations have already raised an alarm on unreliable agents, and once the agents will forget to follow the instructions.
thedailycatblog.com
March 28, 2025 at 11:02 PM
To comply with the rules of new approach to Agent Reliability, Agentspec, Force Agents
Join our daily and weekly newsletters for the latest updates and exclusive content in the industry's leading AI coverage. Learn more AI agents have security and reliability problems. Although the agents allow…
Join our daily and weekly newsletters for the latest updates and exclusive content in the industry's leading AI coverage. Learn more AI agents have security and reliability problems. Although the agents allow…
remains computationally lightweight, with overheads in milliseconds. By combining interpretability, modularity, and efficiency, AgentSpec provides a practical and scalable solution for enforcing LLM agent safety across diverse applications. We also [6/8 of https://arxiv.org/abs/2503.18666v1]
March 25, 2025 at 5:56 AM
remains computationally lightweight, with overheads in milliseconds. By combining interpretability, modularity, and efficiency, AgentSpec provides a practical and scalable solution for enforcing LLM agent safety across diverse applications. We also [6/8 of https://arxiv.org/abs/2503.18666v1]
AgentSpec successfully prevents unsafe executions in over 90% of code agent cases, eliminates all hazardous actions in embodied agent tasks, and enforces 100% compliance by autonomous vehicles (AVs). Despite its strong safety guarantees, AgentSpec [5/8 of https://arxiv.org/abs/2503.18666v1]
March 25, 2025 at 5:56 AM
AgentSpec successfully prevents unsafe executions in over 90% of code agent cases, eliminates all hazardous actions in embodied agent tasks, and enforces 100% compliance by autonomous vehicles (AVs). Despite its strong safety guarantees, AgentSpec [5/8 of https://arxiv.org/abs/2503.18666v1]
agents operate within predefined safety boundaries. We implement AgentSpec across multiple domains, including code execution, embodied agents, and autonomous driving, demonstrating its adaptability and effectiveness. Our evaluation shows that [4/8 of https://arxiv.org/abs/2503.18666v1]
March 25, 2025 at 5:56 AM
agents operate within predefined safety boundaries. We implement AgentSpec across multiple domains, including code execution, embodied agents, and autonomous driving, demonstrating its adaptability and effectiveness. Our evaluation shows that [4/8 of https://arxiv.org/abs/2503.18666v1]
we propose AgentSpec, a lightweight domain-specific language for specifying and enforcing runtime constraints on LLM agents. With AgentSpec, users define structured rules that incorporate triggers, predicates, and enforcement mechanisms, ensuring [3/8 of https://arxiv.org/abs/2503.18666v1]
March 25, 2025 at 5:56 AM
we propose AgentSpec, a lightweight domain-specific language for specifying and enforcing runtime constraints on LLM agents. With AgentSpec, users define structured rules that incorporate triggers, predicates, and enforcement mechanisms, ensuring [3/8 of https://arxiv.org/abs/2503.18666v1]
Haoyu Wang, Christopher M. Poskitt, Jun Sun: AgentSpec: Customizable Runtime Enforcement for Safe and Reliable LLM Agents https://arxiv.org/abs/2503.18666 https://arxiv.org/pdf/2503.18666 https://arxiv.org/html/2503.18666
March 25, 2025 at 5:56 AM
Haoyu Wang, Christopher M. Poskitt, Jun Sun: AgentSpec: Customizable Runtime Enforcement for Safe and Reliable LLM Agents https://arxiv.org/abs/2503.18666 https://arxiv.org/pdf/2503.18666 https://arxiv.org/html/2503.18666
