SABER: Small Actions, Big Errors — Safeguarding Mutating Steps in LLM Agents

Despite rapid progress in LLM agents, performance on long-horizon, tool-using tasks remains fragile. To better understand this fragility, we ask a simple question: do all actions contribute equally to failure? Analyzing execution traces on $\tau$-Bench (Airline/Retail) and SWE-Bench Verified, we decompose trajectories into \emph{mutating} (environment-changing) vs non-mutating steps and formalize decisive deviations—earliest action-level divergences that flip success to failure. A logistic regression reveals that each additional deviation in a mutating action reduces the odds of success by up to 92\% on Airline and up to 96\% on Retail for SoTA models. In contrast, deviations in non-mutating actions have little to no effect. Errors also grow with context length as agents drift from role and act on stale constraints. Motivated by these observations, we introduce SABER, a model-agnostic, gradient-free, test-time safeguard that (i) adds mutation-gated verification, (ii) injects Targeted Reflection before mutating steps, and (iii) performs block-based context cleaning. SABER delivers consistent gains—e.g., Qwen3-Thinking: +28\% relative on Airline, +11\% on Retail, and +7\% on SWE-Bench Verified; Claude: +9\%,+7\%. We further identify ceiling effects in $\tau$-Bench, where annotation errors and underspecified tasks artificially cap model performance. To address this, we release $\tau$-Bench Verified, which restores benchmark headroom through targeted revisions. Our results argue for action-level analysis, targeted safeguards, and reliable evaluations as prerequisites for robust multi-turn agents.