UndoBench: Separating Task Competence from Recovery Capability in Tool-Using AI Agents
cs.SE, cs.AI
Submitted: 2026-10-04
Updated: 2026-10-04
Terminology
Sources
- ReliabilityBench: Evaluating LLM Agent Reliability Under Production-Like Stress Conditions
- OccuBench: Evaluating AI Agents on Real-World Professional Tasks via Language Environment Simulation
- ContainmentBench: Trace-Based Evaluation of Post-Exposure Containment in Tool-Using LLM Agents
- Where Does Exactly-Once Live? Model, Harness, and Tool-Contract Effects on Duplicate Side Effects in LLM Agents
- MemTX: Transactional Belief Commit for Stateful Agent Memory
- Verified Tool Calls Improve LLM Agent Reliability Under Non-Atomic Failures
- Atomix: Timely, Transactional Tool Use for Reliable Agentic Workflows
- EvoUndo: Recoverability-Constrained Self-Evolution for LLM Agent Harnesses
- The Verifier Tax: Horizon Dependent Safety Success Tradeoffs in Tool Using LLM Agents
- ToolMisuseBench: An Offline Deterministic Benchmark for Tool Misuse and Recovery in Agentic Systems
Related papers
- Falsification-Based Verification of LLM-Generated Optimization Models: Sound Test Batteries and Their Detection Limits
- GitSkills: A Dataset of Agent Skills on GitHub
- SABER: Benchmarking Operational Safety of LLM Coding Agents in Stateful Project Workspaces
- PackMonitor: Enabling Zero Package Hallucinations Through Decoding-Time Monitoring
- IntentCoding: Amplifying User Intent in Code Generation
- Incentives and Outcomes in Bug Bounties