Tag
IBM SIL introduces JevAct, a runtime decision-gate framework that intercepts tool-using LLM agent actions, verifies policy compliance via compiled code checks and a prefill-only decision model, and feeds back corrections—boosting Opus 4.6's τ²-bench Telecom pass^3 success from 39.5% to 80.7% without modifying models or tools.
论文提出 CASEL（Consequence-Aware Selective Execution Layer），一种模型无关的运行时分配策略，通过综合校准失败概率、后果严重性、可逆性、歧义性和证据新鲜度来为工具调用选择执行、验证、沙箱、澄清、审批或放弃等模式，仿真结果显示其将有害副作用轨迹比例从 14.75% 降至 7.02%。
MemMux is a local runtime that treats resource governance for fleets of parallel coding agents as runtime verification, providing per-agent memory attribution, complete reclamation, escaped-process visibility, and bounded footprint. Benchmarks show it keeps a 7.5 GiB budget under zero swap and detects all injected escaped processes, with an open-sourced engine and benchmark.
AWS introduces Dogwood, an open-source governance language for AI agents that uses temporal logic to enforce policies over sequences of tool calls, enhancing runtime verification.
TRUSS is an evidence-guided framework for generating functionally effective and safety-reliable Agent Skills, improving task effectiveness from 17.11% to 52.94% and security rate to 100%.
SemaPLC is a verification-gated agent harness for PLC code generation that uses external compilation and live runtime execution to achieve higher verified pass rates than baseline methods.
This paper introduces explanation-based runtime verification for ML-driven optical networks, using model explanations to assess the soundness of individual decisions before execution in the network control loop, demonstrating effectiveness in intercepting erroneous decisions while preserving high automation rates.
This paper presents EntropyRuntime, a discrete-time control system for single and multi-agent LLM-driven and robotic agents that uses five execution gears with utility-gated dispatch and event-driven fallback to ensure safety, stability, and continuity. It provides formal proofs and evaluates on a three-agent UR5 robotic assembly cell, achieving 99.6% anomaly detection rate.