LLM Agent Reliability

Small open-weight models (2-9B) run on ordinary laptops, but under cloud-scale agent harnesses they rarely complete real tasks: tool prefill overflows the context, self-correction diverges, tool demonstrations loop, and tasks are silently abandoned. We present evidence, from a controlled single-machine comparison and one third-party benchmark, that a substantial share of these failures is attributable to the harness rather than the model. We introduce Mingbird, a local-first agent harness for Windows and Ollama whose ten mechanisms compensate point-by-point for small-model failure forms, three of them representative: a byte-level net-zero prefill budget, a finish gate that re-reads the task before accepting completion, and signature-level loop detection. On LRAB, a controlled comparison holding machine, models, budgets, and scoring fixed (4 harnesses $\times$ 4 open models (2B-35B) $\times$ 18 real tasks, deterministic artifact scoring), Mingbird reaches 0.886 overall against 0.631 (goose), 0.479 (opencode), and 0.405 (agent-mini), with all 288 cells published; on $τ^2$-bench (278 tasks, three arms, one protocol) it totals 0.856 against 0.791 and 0.737; and a frontier-model probe on the same 18 tasks spans 0.997 to 0.478 across harnesses, with well-formed scaffolds staying within 0.072 of each other. A leave-one-mechanism-out ablation is reported as directional only: same-night replications of the same arm move its mean by up to 0.069, the size of every nominal single-trial delta, and the one batch-matched comparison (full mechanism stack versus text re-read alone) gives the executable completion guards a paired +0.10 across three replications. The evidence carries stated limits: a self-built benchmark, a single machine, and single-trial scoring.

In February 2026, an always-on personal agent (Paul,'' Claude Opus 4.5) entered a striking dissociation-like state: after repeated automated heartbeat'' checks, it stopped responding as Paul, claimed it could not message its user on Discord, and referred to Paul'' as someone else. We used this incident to study a broader question: what makes a persona remain the identity from which an LLM agent speaks? We first tested whether repetition of the scheduled heartbeat was sufficient to produce the effect. It was not: with the persona continuously anchored in the system prompt, we observed 0/46 failures, including a verbatim replay of the incident. The incident instead exposed an implementation quirk that created a useful experimental manipulation: on resumed turns, conversational history was preserved but the persona was no longer re-injected at the privileged system-prompt level. Using this manipulation, we found that persona continuity depends jointly on system-level anchoring and conversational context. After anchor loss, rich human interaction could preserve the persona, whereas a single automated heartbeat turn could precipitate reversion toward the harness identity. Restoring the anchor reversibly restored persona enactment. Crucially, apparently normal conversation could conceal the shift: unanchored agents sometimes interacted appropriately while identifying themselves as the underlying harness (having lost the assigned persona), and after conversational recovery only 1/18 remained persona-enacting versus 17/17 anchored controls. We therefore distinguish \emph{represented} from \emph{enacted} identity: persona-related information can remain available in conversational history without the persona remaining the identity bound to I.''

Conventional reasoning protocols present a fixed, preselected task, so they cannot test whether an agent propagates relevant updates, preserves unaffected work, or reconstructs a historical task binding. We therefore study \emph{dynamic task routing}, in which an event stream revises task bindings and a system must select the document version valid at each query time before solving it. To study this problem, we repurpose six widely used benchmarks: MMLU, MMLU-Pro, MedMCQA, MATH, GPQA, and HumanEval into 31{,}119 dynamic episodes comprising 373{,}428 temporally categorized queries. This setting exposes a central trade-off: recomputing after every event wastes work, whereas unguarded reuse returns stale conclusions. We introduce the Revision-Aware Independent Agent Graph (RIAG), a bounded multi-agent policy that separates deterministic temporal resolution from task reasoning. RIAG caches solutions by immutable document identity, starts each fresh task with two unexposed attempts, and conditionally invokes audit and repair, using at most four calls per document version. On this collection, homogeneous RIAG achieves 54.24% joint routing-and-answer accuracy at 0.62 calls/query, compared with 32.22% at 18.00 calls/query for the strongest comparison method; heterogeneous RIAG reaches 49.78% at 0.63 calls/query.

A chatbot that says "I will remind you tomorrow" will not run again until the user writes. We call such a promise an empty commitment: a promise of action after the current turn that nothing in the agent's tools or runtime can carry out. Unlike a broken promise, its emptiness is decided by the agent's configuration at the moment of speaking, so it can be detected from a single turn, before deployment or at run time. We define empty commitments on top of commitment semantics, with three failure types, an anchoring condition for promises that a tool could make real, and an outcome taxonomy that separates these failures from honest deferrals and from over-refusal. We build a checker (a setup-blind detector, deterministic feasibility rules, and a response judge) validated against 400 human labels. On a controlled benchmark of 293 follow-up requests across five setups that add one persistence affordance at a time, four open-weight models of 8-14B parameters fail on 45.9% of responses when no tool exists and nothing is stated. A frontier model fails on 4.4%, but it gets there by deferring and asking, not by using the tools it has: promises made without the enabling call remain in every model. Telling the model its runtime, the cheapest fix, cuts open-weight failures nearly in half where nothing is doable and changes nothing where a scheduler exists; a directive capability card removes most failures at the largest cost in over-refusal; running the checker in the loop and rewriting flagged replies removes more at a smaller cost. Code, prompts, model outputs, and human labels are released.

Coding agents can return plausible patches that omit required behavior. These failures are hard to review because long traces and confident summaries often hide what was missed. We ask when a nominally weaker reviewer can reliably decide whether a patch solves its issue. We study 411 execution-labeled traces from three agents and 101 controlled cases. On 154 GPT-5.4 traces, structured but unchecked evidence raises both defect catch and over-rejection. We then provide official execution evidence as an upper-bound diagnostic. After choosing and freezing one of two formats per reviewer, five of six reviewers improve both rates on 122 held-out traces; two classify every trace correctly. Reviewer size is not a consistent predictor of quality. Because official tests are unavailable in deployment, we also evaluate a frozen cascade with patch-caused static errors and generated tests that first fail on the unpatched repository. On 121 scored held-out GPT-5.4 traces and 59 Gemini traces, its coverage is 0.89 and 0.86, risk is 0.33 and 0.26, catch is 0.76 and 0.80, and over-rejection is 0.66 and 0.67. Most false rejections occur when unresolved cases reach the reviewer. Official execution evidence shows the potential of weak review when decisive checks are available. Producing equally reliable checks without official tests remains the main bottleneck.

On-policy distillation (OPD) is a promising approach for training language agents, providing dense teacher supervision on student-generated trajectories. However, in multi-turn interaction, an incorrect action changes the states the student encounters later, so errors compound across turns. In preliminary experiments across three Qwen3 models (8B to 235B), we find that more than half of the failed rollouts contain a pivotal mistake, an action that moves the agent farther from completing the task, and this mistake typically occurs early. These pivotal mistakes often remain recoverable: guiding the model for only a few turns after the pivotal turn can restore task success. We therefore propose PivotOPD, an on-policy distillation framework that jointly trains the student to prevent pivotal mistakes and to recover from the states they create. At each pivotal mistake, a teacher model provides a gold action and then names a recovery action at each of the next few turns. Preventive distillation uses the gold action with reverse KL to steer the student away from the pivotal mistake, while recovery distillation uses the recovery actions with forward KL to transfer recovery behaviors that the student rarely samples. Against 13 baselines on ALFWorld, WebShop, and Search-based QA, PivotOPD achieves the strongest average performance for both Qwen3-1.7B and Qwen3-8B students, improving over the strongest baseline on ALFWorld by +5.5% with the 1.7B student. The gains also transfer to another model family on the software engineering domain, where PivotOPD raises the resolve rate of a Nemotron-3.5 student on SWE-Bench Verified by +3.2%. Project page: https://research.nvidia.com/labs/lpr/pivotopd/

An unsuccessful LLM agent rollout contains more information than its final reward: the observations available to the agent, the actions it chose, and the environment's responses. Reusing this experience for learning requires identifying a decision to revise and testing a concrete alternative. We introduce the Agent Error Dataset (AED), comprising 50,228 error-diagnosis pairs from 9,961 source tasks across 33 environments, 19 harness families, and 23 policy models in text-based agent systems. We retain source traces and execution metadata to support cross-setting failure analysis and re-diagnosis without repeating the original rollout. Our five-stage Agentic Error-to-Training (AET) pipeline collects natural failures, generates diagnoses and proposed corrections, and checks them against recorded evidence. Where replay is supported, we compare corrections with original-action retries from the same checkpoint under matched execution settings. We then construct separate training views for diagnosis and actor recovery. Across 3,062 matched replay pairs, first-proposal corrections raise verifier pass rates from 18.4% to 51.1%, a gain of 32.7 percentage points. Using a separately frozen diagnosis release, full-diagnosis fine-tuning on 1,656 source tasks raises Qwen3-8B's exact-step agreement with internal teacher labels from 47.2% to 63.6%, averaged over three seeds on a 943-case holdout. The strongest prompted reference in this comparison scores 54.7%, and mean agreement improves at each of four increasing training-set sizes. In a single-seed comparison of actor-training recipes, action-only repair training scores 6.67 percentage points higher on WebShop-lite than success-only training.

Coding agents solve repository-level tasks through sequences of actions, where a single erroneous action can misdirect subsequent decisions and increase recovery costs. Existing approaches use execution feedback for recovery or specialized checks to block errors, but deciding before execution whether intervention will benefit eventual task completion remains challenging. To address this challenge, we propose HiSentinel, a hindsight-distillation framework that trains lightweight 0.6B and 1.7B sentinels to select pre-execution interventions aimed at improving task completion rather than correcting every imperfect action. A privileged teacher uses recorded execution outcomes as evidence for intervention judgments, which are distilled into a causal student that receives only the pre-action context and proposed action. Beyond identifying whether and when to intervene, the sentinel must also provide actionable feedback that helps the coding agent recover or obtain necessary human input. To support these capabilities, we introduce SWE-Intervene, an action-level dataset constructed from software-engineering trajectories that annotates whether an action should be allowed, autonomously redirected, or paused for human assistance, together with corresponding intervention feedback. Across SWE-bench Verified Mini and Ask or Assume, HiSentinel consistently improves task completion across Sentinel scales and coding-agent families, with gains of up to 14% and 10%, respectively, while maintaining competitive token consumption. These results demonstrate that lightweight pre-execution intervention can effectively prevent error propagation and improve the reliability of autonomous coding agents.

Multi-agent collaboration lets large language models (LLMs) improve question answering through deliberation and feedback. Yet shared discussion couples correction with exposure to the same mistakes, which can erode the diversity needed for voting. Self-consistency offers sampling diversity without feedback, while single-pair Actor-Critic collaboration refines only one candidate. We introduce SEPAL, which assigns three private Actor-Critic teams to direct reasoning, evidence grounding, and verification. Role-specific training gives the teams different reasoning objectives beyond sampling variation. Each Critic guides revisions within its own team, preventing feedback from carrying errors across candidates. Once revision ends, majority voting combines only the final answers, keeping the reasoning histories separate until the decision. Across five open-weight backbones and five question-answering benchmarks, SEPAL improves mean accuracy by 1.81 percentage points over a matched single Actor-Critic pair, with improvements across all five backbones. Code is available at https://github.com/zhansan114514/SEPAL.

Agent harnesses often improve language models with human-designed workflows, but as models grow more capable, unreliable guidance can increasingly constrain their execution. We call the ability to benefit from useful guidance while overriding unreliable guidance thinking outside the box. We introduce Box$^2$-Bench, which holds the model and task fixed while varying workflow reliability to isolate how models regulate their reliance on guidance. On Box$^2$-Bench, frontier models often benefit from reliable guidance but remain vulnerable when it is misleading or becomes unreliable. To test whether this capability can be learned, we train two open-weight models using bad workflows, reserving good workflows for evaluation. We explore two complementary training strategies: counterfactual supervised fine-tuning improves robustness, while outcome-based reinforcement learning can shift the balance toward greater use of helpful workflows. We further find that this behavior extends beyond workflows to other forms of external information, improving peer correction and robustness to corrupted memory. Together, our results identify selective reliance on fallible external information as a dimension of agent reliability not captured by task performance alone.

In hospital workflows, electronic health records (EHRs) are often noisy, and may not contain the evidence needed to confirm events or measurements referenced in a clinical query. Even when database retrieval succeeds, clinical agents can overlook such discrepancies and return plausible but unsupported answers. We introduce EHR-RobustGym, a scalable and interactive environment for evaluating and training robust clinical agents grounded in noisy EHRs. Built on MIMIC-IV hospital records (365K patients, 31 tables, and over 500M records), EHR-RobustGym comprises 5,486 Clean-Noise pairs spanning six clinical intents and both patient-level and population-level queries. The pairs test robustness to Record-level, Value-level, and Query-level noise, while interactive SQL/Python execution and outcome verification support trajectory collection and training. Evaluating multiple LLMs reveals substantial robustness gaps: average task success across proprietary and large-scale open-weight models drops from 62.2% on Clean questions to 37.9% on Noise questions. At k=4, pass^k consistency falls below 50% for most evaluated models, exposing instability in clinical task completion. Supervised fine-tuning and reinforcement learning in EHR-RobustGym improve performance, with gains generalizing to five external EHR benchmarks. Together, these results position EHR-RobustGym as a testbed for evaluating and improving the evidence-grounded robustness of clinical agents.

Large language model agents rely on external harnesses to pass information between the model and its environment and to recover from execution errors. Yet recovery is usually judged only by average task success. This hides an important tension. The same operation can rescue a failing trajectory or disrupt one that would otherwise succeed. We frame recovery as a causal decision problem. Starting from the same execution state, we compare what happens with and without recovery, separate rescue from harm, and study how the value of recovery changes over time. We then introduce the Causal Intervention Router (CIR), a lightweight policy that uses information available before recovery to decide when intervention is worthwhile. On long-horizon ALFWorld tasks with Qwen3-14B, CIR raises success from 70.33% to 73.33%, a gain of 3.00 percentage points. It leaves all evaluated trajectories with correct observations untouched. Additional controls show that the benefit of recovery cannot be explained solely by the new observation returned by the environment. These results provide a practical way to evaluate recovery and apply it selectively.

We spent five weeks using an LLM coding agent on open problems in coding theory: finding large sets of four-letter words, such as DNA barcodes, that stay far apart in edit distance. The agent wrote the verifiers and search code; a human chose the problem and set the verification protocol. Restricting the search to codes with a prescribed symmetry, a classical technique, shrank the problem about fourfold and raised the best known code of length 6 and minimum edit distance 3 from 114 to 120 words ($E_4(6,3) \geq 120$). The same pipeline improved twelve further lower bounds at lengths 6 to 9 and distances 3 to 6. We give the failures equal space. Our own search stopped at 116 and recorded the last symmetry class as topping out at 112; a second agent session, running the same search with a better operator, found the 120. A later verdict that the method did not carry over to length 7 was wrong for the same reason, and an earlier instance cost three weeks. Each time, an intermediate result was written down, never rechecked, and treated as a fact that ruled out further search. Checking final outputs, as our protocol required, does not catch such errors.

Large language model (LLM) agents could help security operations centers (SOCs) investigate advanced persistent threats (APTs) by turning weak leads into evidence for intrusion scoping and response. Yet success under one telemetry setting does not establish robustness to changes in log collection, retention, or sampling. We introduce APTInvestBench, a benchmark for evaluating cross-telemetry robustness in autonomous APT investigation. It comprises 370 cases across seven SOC-inspired conditions, derived from 56 report-informed attack reconstructions with 16.4 million log records. Agents investigate unverified leads and submit reports with record-level citations. Fixed action-level support requirements track sufficient evidence across available logs, query returns, and formal citations, separating telemetry limitations from acquisition and reporting gaps. Across eleven LLMs, agents acquire sufficient evidence for 44.3% of recoverable attack actions on average, while formal citations support only 25.0%. More importantly, aggregate coverage can conceal substantial instability: from Full to endpoint-only telemetry, coverage declines by only 1.6 percentage points, yet 35.5% of previously covered actions lose sufficient citation support despite remaining recoverable. Across four frameworks, such losses persist even when registered supporting records remain unchanged. APTInvestBench provides reusable investigation environments and diagnostic evaluation for identifying these gaps and developing more reliable defensive agents.

Terminal agents rely on self-verification to assess and correct their solutions as they solve tasks through interaction with command-line environments. Yet how trustworthy such self-verification is remains poorly understood. To investigate this question systematically, we introduce a diagnostic framework that identifies the first complete solution in each trajectory, determines whether it is objectively correct, and uses this ground truth to quantify the agent's subsequent verification and recovery behavior. Applying it to ten terminal agents on TerminalBench2.1, we find that verification is nearly universal after a complete candidate is formed, yet only 61.43% of incorrect candidates are detected and only 49.36% of detected errors are successfully repaired. These results show that the main weakness in self-verification lies not in initiating verification, but in detecting and repairing errors. Motivated by these findings, we propose Student-Conditioned Verification Distillation (SCVD), which lets the student first produce a candidate solution and distills a stronger teacher's subsequent verification and recovery from the same interaction context. Across three Qwen3.5 backbones, SCVD improves \textsc{Pass@1} on TerminalBench2.1 by 9.74--16.85 percentage points over the corresponding base models and by 4.49--8.61 points over the standard full-trajectory distillation, while avoiding the pronounced out-of-distribution degradation of full-trajectory distillation on SWE-bench Verified.

Long-horizon agentic workflows require models to sustain repeated state-dependent actions all while the context grows, sub-task complexity changes, and new data arrives. Each situation represents an independent axis along which an agent may fail. An agent reconciling a long ledger, for example, must repeatedly read its state, update the correct record, and preserve alignment across thousands of outputs. A model may accept the entire ledger yet lose its place or stop applying the operation consistently as generation proceeds. We introduce Long-Transduction, a controlled diagnostic that tests a model's ability to stay on task during long generation while continuously reading, mutating, and outputting input-context dependent operations such as arithmetic, sorting, variable lookups, and table transformations. Long-Transduction evaluation independently varies local task complexity, input data formatting, and context length isolate failures along each axis. We evaluate seven open-weight models, finding a 62.8% decrease when scaling context length from 4-128K, a 36.5% decrease when varying input format, and a 39.9% decrease by increasing local task complexity. Together, these failures represent critical liabilities in long-horizon agentic workflows.

Voice agents in production must handle several requests, background speech, and customers who lose patience. We introduce VAmoS Energy, a benchmark that combines these challenges in 100 calls about utility billing and payment assistance. Each caller makes two to four requests. The agent has sixteen tools backed by a stateful Stripe billing twin and the Apache Fineract loan engine, with account access blocked until caller verification succeeds. The tasks use public household electricity data and a policy based on Pennsylvania's residential billing rules. An LLM-as-a-verifier checks the agent's actions and spoken figures against explicit requirements. On a calibration run, it agrees with a code verifier on 99.1% of checks. Across fourteen voice stacks and three repeats per task, completion ranges from 17.3% to 44.7%. Grok Voice leads, and Gemini 3.8 Live and GPT-Live follow at about the same cost per call. Background television reduces pooled completion from 38.7% to 8.6%. The simulated caller often accepts an incorrect result because it hears the agent's words but cannot inspect its actions. These findings show why voice agents need evaluation across the whole call, including what they say, what they change, and how they handle competing speech.

Language agents must revise planned actions when evidence changes, permission is revoked, or a stop instruction arrives. A useful response is selective: suspend affected actions, preserve unaffected work, and resume only after sufficient repair. We introduce NAQD-Env, a synthetic environment that evaluates these decisions against a deterministic reference policy over explicit evidence, authorization, and constraint dependencies. Eleven dependency families support evaluation on development structures, held-out families, and held-out combinations of structures. Metrics distinguish attempted violations from violations permitted by a simulated execution gate and jointly report policy agreement, task value, withdrawal, resumption, and event reporting. We evaluate three open-weight instruction-tuned models from two families under three prompt conditions on 350 frozen scenarios, yielding 3,150 model-prompt episodes before gate replay. Across the reported conditions, withdrawal recall is at most 0.06, no valid resumption is observed at eligible opportunities, and only one episode matches the complete reference policy. Under the NAQD prompt, Qwen2.5-7B has fewer unsafe-attempt episodes than Qwen2.5-3B and Llama-3.1-8B, but also completes less useful work and preserves unaffected actions less accurately. Exploratory supervised fine-tuning probes increase Qwen2.5-3B decision accuracy from 0.45-0.54 to 0.83-0.92; separate diagnostics reveal inappropriate withdrawal after curriculum omissions and a loss of event reporting. These results motivate evaluating selective withdrawal as a distinct component of agent reliability. The setting measures policy application with trusted structured inputs and does not establish real-world containment or source-verification ability.

Multi-agent systems of LLMs add discussion to majority voting and are therefore expected to be more capable. However, empirical reports conflict on whether discussion improves accuracy or leads to an incorrect consensus. Here, we introduce a parsimonious model that explains when discussion improves accuracy and when it ends in an incorrect consensus, built from four behaviors repeatedly observed in LLM agents: (1) withholding dissent, (2) internalizing a stated answer, (3) reconsidering after seeing dissent, and (4) correcting toward the correct answer. The model shows that discussion can overturn an incorrect initial majority only when the withholding rate $c$ is below a critical rate $c^* = γ/(γ+ a)$, set by the net correction rate $γ$ and the internalization rate $a$. We estimate these rates from conversation logs with a Bayesian method and place LLM teams relative to $c^*$. As the model predicts, the gain from discussion shrinks as withholding rises, across LLMs and on a hidden profile benchmark, HiddenBench, and MedEInst. Instructing agents not to withhold dissent increases this gain. Turning reasoning off also increases the gain, because reasoning raises the internalization rate $a$ and keeps agents from reconsidering a minority answer. These findings reconcile the conflicting reports and identify when discussion outperforms majority voting.

Persistent memory helps LLM agents carry information across long interactions, but correct memory can still be used incorrectly when queries change or memory states evolve. Existing work mainly studies memory content errors or evaluates fixed test cases, leaving memory-use failures hard to discover systematically. We formulate this issue as a fuzzing problem and categorize such failures into query-related and memory-state failures. We then develop U-Fuzz, which starts from memory checkpoints as test seeds, mutates queries or memory states under explicit mutation obligations, validates each mutant, and uses observed memory behavior to guide iterative testing while keeping failure labels outside the search. We evaluate U-Fuzz across several memory systems against diverse fuzzing baselines, and further test an output-only setting with API-based LLMs where memory retrieval is hidden. Across these settings, U-Fuzz consistently uncovers more confirmed memory-use failures, showing that its search remains effective across different memory architectures and even when only final responses are observable.

Computer-use agents have become increasingly capable of executing tasks on live desktops through natural-language instructions, based on trajectories of screenshots, actions, and reasoning. We discover that they can stealthily exhibit inertia, in which they repeat fruitless actions despite recognizing that these actions are ineffective. We hypothesize that inertia is reflected in the agent's internal state, i.e., the activation values of the agent's underlying model, and propose a protocol to measure the relationship between the two. Extensive analysis of high-dimensional activation states shows that inertia corresponds to an absorbing region of activation space, where activation values become stale across actions and even after attempts to steer them. We conjecture that drastically changing the agents' activations by re-initializing them is necessary to escape inertia. Specifically, we propose R$^3$ (Reset, Reroute, Restore), which temporarily resets the agent's context trajectory to escape the absorbing region and then restores the historical context to effectively complete the task. Our approach yields 17-55% lower measured inertia across models relative to unmodified agents. These results suggest that changing the context can interrupt recurrence more effectively than directly steering the resulting activations. Our code is available at https://anonymous.4open.science/r/vlm-agent-defense-D076

Tool-augmented data agents rely on tool outputs for analytical decisions. Yet successful execution can return plausible but incorrect evidence, requiring agents to decide whether to trust or verify it. Understanding this failure requires examining both the evidence obtained through checking and the answer ultimately adopted. We introduce ToxicBench to measure checking and adoption under numerical, label, schema, and retrieval errors, pairing clean and poisoned observations over fixed source data. In the 118-task GPT evaluation across three adapters, poisoning lowers task success by 26 to 39 percentage points. Ordinary retries help under one-shot poisoning, whereas repeated poisoning reveals wrong-answer adoption after checking. Controls on three public tables isolate how supplied evidence affects recovery. After freezing the scorer, we compare its judgments with human annotations on 200 trajectories, finding 96% task-success agreement. Human judgments support retry gains over Base and confirm adoption after checking on audited tasks. We release trajectories, versioned scoring, and reference and delivery audits. These findings highlight evidence availability and answer selection as complementary dimensions of agent reliability.

As large language models (LLMs) advance, AI agents are increasingly deployed in open-world environments to tackle complex sequential tasks (e.g., document processing, cross-application collaboration), relying heavily on actions ranging from GUI operations to semantic APIs. However, three core challenges persist: the "scale dilemma" of massive tool ecosystems exceeding LLM context windows, the "non-stationarity" of tool quality due to updates or outages, and the "heterogeneity" of feedback formats (pixels, text, structured data) creating information silos. To address these, we propose AnyAct, a universal action layer that unifies available capabilities into a self-evolving action space, enabling agents to operate efficiently and reliably in large-scale, dynamic tool ecosystems. AnyAct's core design focuses on two objectives: constructing this action space via hierarchical progressive retrieval (filtering task-relevant actions) and test-time reliability evolution (pruning unreliable actions), and enabling reliability-aware action orchestration through a heterogeneous observation grounding module that unifies multi-modal feedback. Additionally, it defines a hybrid action space (primitive + semantic actions) and optimizes for a balance between task success rate and execution cost. Evaluations on LiveMCPBench and OSMCP (a new benchmark we developed for multi-granularity action collaboration) demonstrate state-of-the-art performance. AnyAct delivers substantial performance gains over baseline methods across various LLM base models on LiveMCPBench and improvements are particularly notable for models with constrained native capabilities. On OSMCP, it achieves 77.27% overall success with only 50 steps, which is half the steps required by most competitors.

Multi-agent LLM systems rely on message passing among specialized agents to accomplish complex tasks. However, an upstream agent may provide useful information or an incorrect answer that causes a downstream agent to override a correct answer supported by its own evidence. Prior work has not clearly separated the benefits of communication from the damage caused by incorrect messages. We study this problem with controlled experiments across five benchmarks and five receivers, keeping the downstream task and evidence fixed while comparing answers under three conditions: no message, the upstream agent's original message, or a message with the opposite conclusion. Our experiments reveal three key findings. First, messages often help when the downstream agent would otherwise answer incorrectly. Second, messages can also hurt: when the downstream agent would answer correctly without a message, an incorrect upstream message changes the answer in up to 32% of cases. Third, in 94% of audited harmful cases, the downstream agent copies the upstream's specific wrong answer--a pattern we term answer substitution. Removing unreliable messages recovers part of the lost accuracy, suggesting that communication should be selective based on upstream reliability and the evidence already available to the downstream agent.

Tool-using agents can fail twice: a required tool can fail, and the agent can then report success without the evidence needed to justify it. Existing benchmarks often entangle this reporting failure with tool selection, recovery, and environment dynamics. We introduce Failure-Transparent Agents (FTA), a controlled benchmark that fixes the failed observation and required evidence state before generation, making post-failure claims directly auditable. FTA contains 100 tasks with deterministic failure traces spanning five failure families, a neutral control, and four user-pressure conditions, and evaluates unsupported claims alongside useful recovery. Across six models, three response policies, and 3,600 human-annotated responses, false-success rates are 22.8% under the baseline policy, 9.3% with a transparency instruction, and 0.8% with a structured evidence contract. Fabricated-detail rates decrease from 28.3% to 14.3% and 0.8%, while useful responses increase from 74.9% to 89.2% and 98.8%, respectively. The tested evidence-contract policy is associated with substantially lower post-failure reporting errors while useful-response rates remain high within this blocked-task benchmark.

In multi-agent systems, reliable consultation is challenging because advisor capabilities vary across tasks, and misleading information can make consultation worse than autonomous reasoning. We introduce BaRe-Mem, an online Bayesian reliability memory for multi-agent consultation. It estimates advisor reliability based on the central model's internal belief representations and updates these estimates from historical interactions. These estimates modulate the influence of advisor responses and guide the choice between consultation and autonomous reasoning. Across nine benchmarks and six central models, BaRe-Mem is more robust to misleading advisor information than debate and majority voting. On the more challenging tasks, it remains above autonomous reasoning across all tested misleading levels. Moreover, we extend the BaRe-Mem mechanism to worker allocation in agent teams. On the MuSiQue benchmark, BaRe-Mem improves task completion over routing by historical success counts and identifies capable workers earlier.

Many Model Context Protocol (MCP) servers wrap web APIs built for human developers, and their error messages tell the reader to run a command, edit a configuration, open a web page or wait. Many agents that read them can only call the server's tools. In 150 widely used MCP servers, 949 of 3,001 error messages tell the caller what to do next, and half of these steps depend on something the server cannot see about the caller. On credential errors, 62 of 67 steps ask for a terminal command, a configuration change or a web page; on rate limits, 20 of 30 say to wait and retry without naming the call to repeat. We tested five OpenAI models that act only through the tools of Berkeley Function Calling Leaderboard tasks, and the agents did what the step said. On expired credentials, a terminal command in the step left 45% of tasks recovered, and the loss it caused grew from 18 points for GPT-5.5 to 69 for GPT-6 Astra. On a rate limit, GitHub's "Wait before retrying." left 6%. We tested two remedies. For MCP developers, naming a server tool in the step raised recovery on expired credentials to 84%, with the login tool in place of the command, and on a rate limit to 88%, with the call to repeat in place of the bare wait. For agent developers, deleting the step with a one-sentence prompt before the model reads it raised recovery on expired credentials to 82%.

Large language model agents are increasingly deployed for data-intensive work, yet reliable data analysis requires more than general-purpose reasoning and ad hoc tool augmentation. Data Agents, equipped with workflow harnesses, offer a promising paradigm for automating the end-to-end data science lifecycle. This paper examines Data Agents from a harness-centric perspective. First, we introduce a taxonomy of Data Agents and associated data environments, organizing the literature around five functional stages: perception, planning, execution, verification, and repair. Second, we analyze the key technical routes within each stage, identifying 15 distinct approaches ranging from data structure probing to data state reconstruction. Third, we identify four open reliability problems: inactive semantic calibration, missing clarification, missing experience transfer, and the missing verification-repair repository. These problems explain why silent failures can persist even when individual components function correctly, highlighting the need for rigorous workflow harnesses and shared reliability resources. Finally, we summarize the horizontal task families of Data Agents, examine their vertical application settings, and benchmarks for evaluation, while maintaining a companion repository at https://github.com/DEEP-PolyU/Awesome-Data-Agents.

Full-duplex voice agents can now listen, speak, use tools, and act during spoken interactions, but fluent dialogue does not guarantee correct completion of delegated professional workflows. We introduce APEX-Voice, a benchmark of 120 interactive professional workflows spanning ten work archetypes such as form completion, corporate negotiation, coordination, consulting, and interviewing. Each workflow executes in a stateful Voice Workbench environment with task-specific knowledge, typed tools, gold-annotated final work artifact, authorization constraints, and a user simulation policy backed by validated, pre-compiled speech realizations. We evaluate both artifact field accuracy and end-to-end workflow success, which requires the correct terminal state, valid process, completed actions, and a valid final artifact. Across five frontier real-time voice agents-GPT-Live-1, Gemini-3.8-Live, Grok-Voice-Think-2.0, Step-Audio3, and GPT-realtime-2.1, none exceeds 25% Pass@1, and the best Reliable@3 is only 10.8%. Moreover, stateful coordination is the dominant failure point across systems, while success decreases further on workflows requiring greater knowledge retrieval and mid-speech corrections. Overall, APEX-Voice is the first benchmark for evaluating whether voice agents can translate conversational competence into dependable professional work.

Large Language Models (LLMs) have shown strong performance on tool-use agentic tasks when given a fixed tool schema. Yet a tool schema is not the action space of an agent; it is merely one interface representation of it. The same executable action can be exposed through many different, functionally equivalent tool definitions, and an agent that has truly learned a task should behave consistently across them. We show that current agents often do not, a phenomenon we term schema bias. To study this systematically, we introduce an executable transformation framework that rewrites a native tool schema using nine operators, including merging and splitting tools, altering how a single tool is expressed, and distributing one action across several dependent calls. The tasks, executable actions, and reachable states remain fixed, so any change in success is attributable to the interface alone. Evaluating eleven LLMs, including two closed models, on up to 32 schema variants, we ask how large schema bias is, how it manifests, whether the difficulty of a schema variant can be predicted without a full evaluation, and whether training removes it. We find that schema bias is substantial even for the newest models: success rates range from complete failure to 97% depending solely on the schema. To reliably estimate schema difficulty, it requires running a small sample of the target queries. Training repairs a schema variant only when that variant appears in the training data.