\system{} combines five gears with utility-gated dispatch for safety in autonomous agents, proving stability for single agents and providing distributed guarantees for multi-agent CPS, evaluated on UR5 robots.
hub
Towards verifiably safe tool use for llm agents
13 Pith papers cite this work. Polarity classification is still indexing.
13
Pith papers citing it
hub tools
citation-role summary
background 2
method 1
citation-polarity summary
years
2026 13representative citing papers
Boiling the Frog is a new stateful multi-turn benchmark that finds an aggregate 44.4% strict attack success rate for incremental safety violations across nine AI models, with rates ranging from 20.5% to 92.9%.
SkillGuard extracts executable environment contracts from LLM skill documents to detect only relevant drifts, reporting zero false positives on 599 cases, 100% precision in known-drift tests, and raising one-round repair success from 10% to 78%.
Proposes an OS-centered privacy framework for on-device AI that treats privacy as institutional accountability, including a threat model, six-part risk taxonomy, privacy-by-architecture controls, and four-level audit rubric demonstrated on Apple, Android, and Microsoft systems.
Symbolic guardrails enforce about 74% of agent security and safety requirements on three benchmarks with mostly simple checks, improving safety without sacrificing utility.
ATLAS-RTC raises first-attempt success on structured LLM generation and tool calling by 20-37.8 points through closed-loop token-level interventions.
No existing AI security framework covers a majority of the 193 identified multi-agent system threats in any category, with OWASP Agentic Security Initiative achieving the highest overall coverage at 65.3%.
AI agents can generate code in a capability-safe Scala dialect that statically prevents information leakage and malicious side effects while preserving task performance.
TRACES learns prefix-level trajectory risk states from LLM hidden representations using weak trajectory-level supervision to enable proactive safety auditing for multi-turn agents.
LLM agent security is reframed as an agent-human interaction issue, supported by a survey showing industry preference for human-centric mechanisms over academic favorites and proposing a new research agenda.
Sovereign Agentic Loops decouple LLM reasoning from execution by emitting validated intents through a control plane with obfuscation and evidence chains, blocking 93% of unsafe actions in a cloud prototype while adding 12.4 ms latency.
Proposes a trust schema including verification levels and a biconditional correctness criterion to verify skills in human-in-the-loop agent runtimes, reducing the need for constant oversight.
citing papers explorer
-
Managed Autonomy at Runtime: Gear-Based Safety and Governance for Single- and Multi-Agent Cyber-Physical Systems
\system{} combines five gears with utility-gated dispatch for safety in autonomous agents, proving stability for single agents and providing distributed guarantees for multi-agent CPS, evaluated on UR5 robots.
-
Boiling the Frog: A Multi-Turn Benchmark for Agentic Safety
Boiling the Frog is a new stateful multi-turn benchmark that finds an aggregate 44.4% strict attack success rate for incremental safety violations across nine AI models, with rates ranging from 20.5% to 92.9%.
-
Skill Drift Is Contract Violation: Proactive Maintenance for LLM Agent Skill Libraries
SkillGuard extracts executable environment contracts from LLM skill documents to detect only relevant drifts, reporting zero false positives on 599 cases, 100% precision in known-drift tests, and raising one-round repair success from 10% to 78%.
-
Local Is Not a Sufficient Privacy Boundary: Governing OS-Integrated On-Device AI
Proposes an OS-centered privacy framework for on-device AI that treats privacy as institutional accountability, including a threat model, six-part risk taxonomy, privacy-by-architecture controls, and four-level audit rubric demonstrated on Apple, Android, and Microsoft systems.
-
Don't Make Models Guess Security and Safety: Symbolic Guardrails for Domain-Specific AI Agents
Symbolic guardrails enforce about 74% of agent security and safety requirements on three benchmarks with mostly simple checks, improving safety without sacrificing utility.
-
ATLAS-RTC: Closing the Loop on LLM Agent Output with Token-Level Runtime Control
ATLAS-RTC raises first-attempt success on structured LLM generation and tool calling by 20-37.8 points through closed-loop token-level interventions.
-
Security Considerations for Multi-agent Systems
No existing AI security framework covers a majority of the 193 identified multi-agent system threats in any category, with OWASP Agentic Security Initiative achieving the highest overall coverage at 65.3%.
-
Tracking Capabilities for Safer Agents
AI agents can generate code in a capability-safe Scala dialect that statically prevents information leakage and malicious side effects while preserving task performance.
-
TRACES: Proactive Safety Auditing for Multi-Turn LLM Agents via Trajectory-State Modeling
TRACES learns prefix-level trajectory risk states from LLM hidden representations using weak trajectory-level supervision to enable proactive safety auditing for multi-turn agents.
-
Reframing LLM Agent Security as an Agent-Human Interaction Problem
LLM agent security is reframed as an agent-human interaction issue, supported by a survey showing industry preference for human-centric mechanisms over academic favorites and proposing a new research agenda.
-
Sovereign Agentic Loops: Decoupling AI Reasoning from Execution in Real-World Systems
Sovereign Agentic Loops decouple LLM reasoning from execution by emitting validated intents through a control plane with obfuscation and evidence chains, blocking 93% of unsafe actions in a cloud prototype while adding 12.4 ms latency.
-
Skills as Verifiable Artifacts: A Trust Schema and a Biconditional Correctness Criterion for Human-in-the-Loop Agent Runtimes
Proposes a trust schema including verification levels and a biconditional correctness criterion to verify skills in human-in-the-loop agent runtimes, reducing the need for constant oversight.
- ForesightSafety-SAGE:A Fully Automated Scenario Generation and Safety Evaluation Framework for LLM Agents