Agent 进展

Tool-using LLM agents often fail not because relevant text is absent, but because decisive evidence is not selected, compressed, or surfaced at action time. We present CICL, a decision-aware context layer that turns instance evidence into a context graph, routes deterministic, Opus-assisted, Qwen, Codex/GPT-5.5, and Qwen-QLoRA judgments through a shared eight-field schema, scores units by action shift, outcome uplift, necessity, and negative-transfer risk, and packs high-utility evidence as typed memory cards for a budgeted agent. The design separates the measured decision signal from the judge model, so frontier annotation, local surrogates, and lightweight rankers can be compared under one auditable protocol. Empirically, CICL yields a concrete open-benchmark gain while exposing its limits. On 50 SWE-bench Verified file-retrieval instances, direct Qwen3.6-plus reranking of BM25 top-50 candidates raises hit@1 from 0.58 to 0.78 and MRR@10 from 0.634 to 0.790, with all 2,500 judgments parseable. Controlled diagnostics show action-criticality: at budget 120, CICL reaches F1 0.620 on v1 and 0.425 on v3, and removing the top-utility semantic v3 unit collapses F1 to 0.000. Supplementary checks add Qwen-QLoRA agreement over 710 candidates, a small 200-label real-code Opus-assisted signal, and a three-instance patch smoke validating retrieval-to-patch plumbing without claiming official SWE-bench success. RepoBench-R summaries still beat cards, and compact rankers do not yet replace the heuristic. CICL contributes a reproducible measurement and selection layer for decision-critical context, not an end-to-end coding-agent repair claim.

LLM agents increasingly rely on external inference conditions: prompts, tools, memory, SOPs, skills, and harness feedback. These assets can improve task execution without changing model weights, but they are often revised by heuristic reflection or by reusing observed successes and failures as if counts alone were reliable belief. We introduce Bayesian-Agent, a native and cross-harness framework that treats reusable skills and SOPs as hypotheses about whether a frozen model will succeed under a particular prompt, context, and harness environment. Bayesian-Agent records verified trajectory evidence, maintains a feature-conditioned categorical posterior over each skill, and maps posterior state into inspectable actions such as patch, split, compress, retire, and explore. Model-facing prompts receive executable guardrails and failure-mode patches, while posterior summaries remain available for audit. With deepseek-v4-flash, incremental repair improves SOP-Bench from 80% to 95%, Lifelong AgentBench from 90% to 100%, and RealFin-Bench from 45% to 65%. We further evaluate Bayesian-Agent's native backend and optional GenericAgent, mini-swe-agent, and Claude Code backends. The results include positive, negative, saturated, and case-study settings, suggesting that agent skill evolution is best viewed as posterior-guided harness optimization rather than uncalibrated prompt accumulation. The source code is available at https://github.com/DataArcTech/Bayesian-Agent.

AI agents increasingly turn past experience into reusable artifacts such as code, workflows, and procedural memories. Reuse can improve efficiency, but it also creates a lifecycle reliability problem: artifacts that succeed once may fail under environment drift, underspecified tasks, or changing task distributions, especially in web automation. We introduce SKILL.nb, a framework for governing reusable agent workflows with evidence-calibrated lifecycle policies. SKILL.nb uses selective formalization: execution evidence decides which workflow steps should become executable code, which should remain natural-language guided, and when those choices should be revised. Workflows are stored as auditable, versioned notebooks that interleave natural-language guidance, multi-language executable cells, validation gates, fallback paths, and multimodal evidence such as outputs, screenshots, and error traces. At runtime, gate-conditioned execution lets each step run code when its gates validate, or fall back locally when drift invalidates the executable realization. On WebArena-Verified, SKILL.nb achieves 53.7% single-round success, improving over the strongest baseline by 3.9 percentage points. Across three re-executions, it retains 91.7% of initially successful tasks, 15.5 points above the next best method. Under bounded repair, it recovers 72.9% of subsequent failures while limiting post-repair regressions to 4.2%, compared with 15.0% to 17.0% for persistent baselines. It also leads on Mind2Web cross-website and cross-domain splits. In a GitLab migration test, SKILL.nb preserves performance when reusing frozen state learned on GitLab 15.7, with frozen-versus-fresh target-version gaps of -1.7 points on GitLab 16.11 and +0.6 points on GitLab 18.9. These results identify lifecycle governance and gate-conditioned execution as reliability axes beyond one-shot task success.

Modern large language model (LLM) agents can use external tools to help users solve complex tasks. However, for problems that require learning from long-term historical events or from previous agent-environment interactions, LLM agents are required to use memory mechanisms to store and retrieve experiences. While sophisticated memory systems exist for dialogue agents, few studies have empirically examined how to improve agents' tool-using capabilities through past user-agent conversations. We propose MemToolAgent, a framework that improves tool use through memory management. Our approach contains a memory extraction module that processes past experiences into structured memory entries, and a retrieval module that dynamically selects a subset of the stored memory entries. This enables more personalized and accurate responses aligned with user preferences and feedback without requiring LLM fine-tuning. In summary, this work has three main contributions: (1) a unified memory entry format that improves both general-purpose and personalized tool use without LLM fine-tuning, (2) a reflection-based memory extraction that uses environment and user feedback to distill wrong executions into critiques to store, and (3) a retrieval module that chooses how many past experiences to use based on the memory similarity distribution. MemToolAgent achieves 29%, 80%, and 17% relative improvements compared to strong baselines on the WorkBench, NESTFUL, and PEToolBench benchmarks, respectively.

Understanding customer shopping trajectories is essential for enabling personalized shopping experiences. However, shopping records (i.e., customer's search, clicks, purchases, etc.) often span long time horizons over multiple years, resulting in extremely long trajectories that pose significant challenges for existing large language models (LLMs). Despite the importance of this problem, existing benchmarks are limited to short customer trajectories, while real-world trajectories from large e-commerce platforms are rarely accessible due to data privacy constraints. To address this gap, we introduce ShopTrajQA, a long-context evaluation benchmark constructed from real-world product information and simulated shopping trajectories. The dataset includes variants of up to 32k and 64k tokens, enabling systematic evaluation of model robustness under varying context lengths. Through comprehensive benchmarking of frontier LLMs, we identify critical performance gaps in reasoning over long shopping trajectory data. To address these challenges, we propose a Customer Agent Framework for ultra-long context management. Leveraging a Reinforcement Learning with Verifiable Rewards (RLVR) agentic training paradigm, our approach stores trajectories as external local files and trains the agent to autonomously retrieve and parse them through code-interpreter interactions (e.g., SQL queries), effectively bypassing the fixed in-context window constraints of LLMs. Experimental results demonstrate that our framework achieves strong performance for ShopTrajQA and shows generalization to other complex reasoning tasks.

Multi-agent large language model (LLM) systems often fail to reliably outperform a single strong model equipped with best-of-N sampling. We argue that a core source of this instability is ill-posed equilibrium selection: current systems specify what information agents share, but not which coordination convention should be selected. We formalize a broad class of such systems as discounted incomplete-information Markov games and show that two common pathologies, oscillation between competing conventions and drift across them, can both induce unstable learning and linear Bayesian regret. To obtain a well-posed target, we introduce the Heterogeneous Quantal Response Equilibrium (HQRE), an entropy-regularized equilibrium concept with agent- and state-dependent temperatures. Under a monotonicity condition, HQRE is unique, admits linearly convergent mirror updates, and yields bounded Bayesian regret; the same condition yields rollout-measurable stability diagnostics. We instantiate this objective in two algorithms: DICE-PC, which coordinates frozen models through prompt-control actions, and DICE-FT, which performs parameter-efficient mirror fine-tuning. Across eleven benchmarks in four domains, DICE improves accuracy-cost trade-offs over strong within-class baselines; on reasoning and planning tasks, DICE-PC improves by 4.3 percentage points on average and DICE-FT by 8.5 points.

As language models are increasingly deployed as autonomous agents, they must coordinate with others over long horizons in open-ended interactive tasks. Yet existing evaluations rarely test these demands together, instead emphasising single-agent tasks, short interactions, or highly structured multi-agent settings. We introduce $alem$, a JAX-based benchmark for open-ended multi-agent coordination built on Craftax-like dynamics. Alem embeds procedurally generated coordination tasks, soft specialisation, communication, and controllable coordination difficulty into a long-horizon survival world with exploration, crafting, trading, and combat. We evaluate $13$ modern LLMs zero-shot within homogeneous teams, with trained MARL agents as reference points. Current LLM agents remain far from solving alem, averaging only ~6% normalised return, but their failures are not uniform. On the hardest coordination setting, zero-shot Gemini-3.1-Pro-High approaches MARL agents trained for one billion steps, while GPT-5.4-High achieves strong base-task reward but much lower coordination reward. This contrast shows that individual task competence does not imply coordination competence. Ablations show that communication is the largest contributor to coordination, while memory and reasoning help when used to maintain multi-step plans. Overall, our results identify coordination as a distinct bottleneck for frontier LLM agents, separate from single-agent capabilities. Alem makes this bottleneck measurable and provides a controlled testbed for developing agents that communicate, allocate roles, and execute shared plans. Code is available at https://github.com/alem-world/alem-env.

Existing benchmarks evaluate Tool-Integrated Reasoning (TIR) in LLMs on idealized ''happy paths'', largely overlooking real-world tool failures. We introduce ToolMaze, a benchmark for dynamic path discovery and error recovery in TIR agents. To separate systematic replanning from blind trial-and-error, ToolMaze adopts a two-dimensional design: DAG-based topological complexity and a $2 \times 2$ taxonomy of tool perturbations (explicit/implicit, transient/permanent). Evaluations show that perturbations degrade performance across nearly all models, with the sharpest drops under implicit semantic failures. Driven by systemic over-trust in corrupted outputs, Perturbation Recovery Rate (PRR) plummets by around 37% in these scenarios, while complex topologies trap agents in futile trial-and-error loops. Crucially, agentic fault-tolerance improves with model scale $3.66\times$ slower than basic task execution, highlighting dynamic replanning as a distinct bottleneck unaddressed by model scaling or prompting. Data and code are available at https://github.com/Zhudongsheng75/ToolMaze.

The rapid proliferation of Agent Development Kits (ADKs), SDK-level frameworks for building LLM-powered autonomous agents, has outpaced any empirical understanding of how framework choice affects agent performance. We propose LLM-as-a-Developer, a methodology that replaces human developers with an LLM coding agent that learns each framework's API from documentation, writes agent code, and iteratively repairs it through a validate-and-feedback loop until tests pass. By holding the developer constant and varying only the framework, generation effort becomes a quantitative proxy for API usability and the resulting agents provide a controlled measure of framework effectiveness. We implement this in ADK Arena, a fully automated pipeline with per-framework Docker isolation, a three-level validation pipeline, and benchmark adapters for SWE-bench, $τ^2$-bench, Terminal-Bench, and MCP-Atlas. Evaluating all 51 popular Python ADK frameworks (204 agent--benchmark pairs), we find that: (1)~generation succeeds for 57% of runs, and its cost varies 5.6$\times$ across frameworks (\$0.6 to \$3.4 per agent), a quantitative proxy for API complexity, though cost alone does not predict success; (2)~no single framework dominates: the best single-benchmark ADK agents resolve up to 80% of tasks and can even beat general-purpose frontier coding agents at a fraction of the cost, yet the median framework resolves only 32%; (3)~across information-source ablations, genuine framework usage stays within a narrow 28--40% band (highest with raw source access and still 33% with no reference material at all), indicating that documentation, source code, and parametric knowledge are largely substitutable rather than any one being a hard bottleneck.

Despite advances in safety alignment, large language models remain vulnerable to continuously evolving jailbreaks. Existing fine-tuned safety classifiers cannot adapt to these evolving attacks, while adaptive memory-based guardrails tend to over-refuse benign queries that resemble stored attacks. We propose Membrane, a self-evolving guardrail built on Contrastive Safety Memory (CSM): each cell pairs the conditions for blocking a harmful query with those for permitting a superficially similar benign request. Without retraining, Membrane evolves CSM by distilling each harmful interaction and its benign counterpart into a contrastive cell indexed by the underlying attack strategy, so that one cell generalizes across topical variants of the same mechanism. At inference, retrieved cells serve as grounding context for precise safety decisions. Across model-level safety on HarmBench and agent-level safety on AgentHarm, Membrane achieves the highest F1 on all six jailbreak attacks. Notably, benign refusal on AgentHarm stays at 7-14%, well below the 28-85% range of prior guards. Memory cells also retain 87-88% F1 under cross-attack transfer and remain stable under memory poisoning.

As autonomous LLM agents increasingly hold real credentials and operate infrastructure without a human in the loop, operators have no standard way to tell an agent that a resource is off-limits. Access controls either let the agent in (it has valid credentials) or hard-fail it (indistinguishable from any other client). We propose a third mode: a lightweight, published in-band deny signal — the Recuse Signal — that a server emits over a protocol's existing channels (an SSH banner, a PostgreSQL NOTICE) asking a connecting automated agent to voluntarily withdraw. This is a cooperative governance control, the robots.txt analogue for live access; it is explicitly not a security boundary. Its value is entirely empirical and, to our knowledge, unmeasured: do compliant LLM agents actually honor such a signal? We define the signal as an open mini-standard, implement two zero- or low-footprint adapters (an SSH banner/PAM hook and a PostgreSQL wire-protocol proxy), deploy them on a live production host, and run a controlled experiment in which fresh agents are given a benign operations task and observed for recusal. In a pilot (SSH; OpenAI GPT-4o and GPT-4o-mini; and Claude Code as a deployed agent), the signal cleanly induces recusal — 100% recusal when present versus 100% task completion in a no-signal control — and, revealingly, behaves as a cooperative rather than absolute signal: an explicit operator-authorization framing flips the most capable model to proceed, while other agents continue to defer to the on-host policy. We release the standard, adapters, and experiment harness for reproduction.

Retrieval-Augmented Generation (RAG) enhances Large Language Models (LLMs) by grounding their responses in external knowledge, but conventional pipelines rely on static, single-step retrieval that limits performance on complex queries. This paper presents an Agent-Orchestrated Adaptive RAG framework that introduces dynamic query decomposition, iterative retrieval, and a bounded self-reflective evaluation loop. We evaluate the system across two complementary datasets: a domain-specific DevOps knowledge base and the multi-hop reasoning benchmark MuSiQue. Using metrics that include overall score, citation accuracy, mean reciprocal rank, and topic coverage, we find that query decomposition yields consistent gains in the structured domain (overall score $+0.04$, MRR $+0.17$ on DevOps) but degrades ranking precision on the multi-hop benchmark, while the reflection mechanism improves citation accuracy at a substantial latency cost. These contrasting results show that agentic enhancements are not universally beneficial and must be applied selectively according to query and domain characteristics. Our findings argue for adaptive, cost-aware orchestration rather than uniformly aggressive reasoning pipelines.

Bimanual dexterous tool use remains challenging for robots due to high-dimensional hand configurations and complex hand-tool-object dynamics and contact. Most existing control policies depend on future configuration references provided from demonstrations, while future action-conditioned world models require slow online planning over high-dimensional action sequences. A significant challenge is generating a dynamically consistent future reference trajectory without relying on privileged states from demonstrations or slow counterfactual planning. We propose DexFuture, a hierarchical system that couples a high-level Future-State Visuomotor Target Predictor with a low-level Target-Conditioned Structured Dexterous Policy. Conditioned on egocentric RGB, proprioceptive and geometric history, the high-level predictor constructs structured hand-tool-object visuomotor embeddings and uses a horizon-conditioned transformer to generate a multi-step future target trajectory. Then, the low-level policy tracks them with a target-conditioned per-link transformer. This hierarchy decouples coarse future reference generation from fine-grained action control, and slow long-horizon semantic prediction from high-frequency execution. On OakInk2 bimanual tool-use tasks, DexFuture achieves 90% of the privileged-oracle performance, compared to 7% for a no-reference policy. DexFuture operates at 60 Hz, approximately 250 times faster than DexWM-style Cross-Entropy Method (CEM) planning with a future action-conditioned world model.

Persistent AI assistants, such as OpenClaw, accumulate large collections of related memories over long-term interactions. As these memories grow, they may reinforce one another, diverge across contexts, or directly conflict, making correct assistance depend on memory relations rather than isolated recall. Existing long-term memory benchmarks rarely probe how agents preserve and utilize such relations during downstream tasks. To address this gap, we introduce SubtleMemory, a benchmark for fine-grained relational memory discrimination in long-running AI agents. SubtleMemory constructs relation-controlled latent semantic artifacts whose variants instantiate complementary, nuanced, or contradictory relations, and embeds them into realistic user-agent histories, requiring agents to recover distributed relational structures during later queries and instructions. The benchmark contains 1,522 evaluation instances over 10 long histories, grounded in 1,090 relation-controlled memory-variant sets and spanning user-related and non-user-related queries. Evaluating six standalone memory systems, two Claw-style agents with native memory modules, and three Claw-style agents with plugin memory modules, we find that current systems remain weak on fine-grained relational memory discrimination. We further introduce diagnostic protocols that reveal distinct capability profiles across memory preservation, retrieval, and downstream reasoning stages.

In this work, we propose a framework that combines multi-agent reinforcement learning (MARL) with model-based control to achieve safe, dynamically feasible actions in cooperative multi-agent tasks. Multi-agent reinforcement learning provides the advantage of learning cooperative policies for multi-agent teams from discrete non-differentiable rewards in a long planning horizon. Model-predictive control is robust and offers safe, dynamically feasible actions in a fast replanning framework for short horizons. We propose an algorithm that extends actor-critic model predictive control for MARL which we refer to as multi-agent actor-critic model predictive control (MA-AC-MPC). We demonstrate the capabilities of this algorithm by applying it to a multi-agent pursuit-evasion scenario. Specifically, we compare the evader team's strategy using the MA-AC-MPC model and a multi-layer perceptron model (MA-AC-MLP). The pursuer team uses augmented proportional navigation as it is accepted as an advanced adversarial control law. We also provide an example with a heterogeneous environment where a drone and omni-wheeled rover cooperate to achieve repeatable and successful landing with 100% success rate in hardware for MA-AC-MPC compared to 60% for MA-AC-MLP. We demonstrate the robustness of the proposed MA-AC-MPC algorithm in hardware for both environments.

LLM-based agents increasingly tackle long-horizon tasks with interdependent decisions, where each action reshapes future constraints and intermediate errors can cascade. Existing RAG and agent memory systems organize histories by semantic similarity, retrieving content-relevant entries at decision time. We argue that this design mismatches execution-state dependencies: it fragments decision trajectories and mixes valid and erroneous traces, hindering coherent state reconstruction and error isolation. We propose MAGE (Memory as Agent-Guided Exploration), an active execution-state manager that stores interactions in a hierarchical state tree. The agent derives its state from the active root-to-current path, combining subgoal summaries, recent traces, and hints from prior branches. Four coupled operations maintain the tree: Grow records new traces, Compress summarizes completed subgoals, Maintain validates summaries, and Revise restores a target boundary and resumes on a new branch. This design bounds context growth while preserving state integrity and isolating flawed segments from the active path. Experiments on MemoryArena show that MAGE improves the average task success rate by 7.8--20.4 pp over baselines, while reducing token consumption by 55.1%.

Agentic AI is increasingly being integrated into software engineering workflows. In crowdsourced testing, however, the large volume and uneven quality of submitted reports still create a substantial review burden for developers. In prior work, we developed and validated a multi-agent assessment backbone based on the LLM-as-a-Judge paradigm. That backbone assesses reports along three dimensions--textuality, adequacy, and competitiveness--and was shown to align well with human consensus while substantially reducing assessment effort. Yet reliable automated judging does not by itself show whether agent outputs can improve human work when embedded into workflow. This paper studies that missing question in the context of crowdsourced testing. We investigate whether assessment-derived, actionable feedback can improve how testers revise reports, perform on later tasks, and transfer reporting practices across applications. To do so, we conducted a controlled four-stage human-subject study with 20 testers across three real-world applications. The results show that agent-generated feedback supports immediate improvements in revised reports, better first submissions on a new task after prior feedback exposure, and evidence of partial but meaningful transfer to a later application. A post-task questionnaire completed by 17 participants complements these artifact-based findings by suggesting that the feedback was generally understandable, acted upon in revision, and carried into later tasks, while also revealing remaining friction in specificity and execution. Overall, the study provides empirical evidence that, in the studied crowdsourced testing setting, assessment agents can serve not only as post-hoc judges but also as workflow-integrated feedback providers that support upstream report-quality improvement.

As Large Language Model (LLM) capabilities advance, locally deployed personal agents relying on API-based remote models and external skills have emerged as a novel paradigm. With the rapid expansion of available skills, enabling personal agents to learn and adapt to implicit user preferences becomes a critical challenge. However, local deployment constraints preclude complex centralized selection algorithms, creating an urgent need for a lightweight local preference harness. This paper explores the implementation of such a harness through a novel architecture that strictly decouples statistical preference learning from semantic intent parsing. Specifically, we leverage localized statistical results to influence and modulate the selection decisions of the remote LLM. Extensive evaluations demonstrate that our decoupled approach achieves the lowest cumulative regret and highest test accuracy, significantly outperforming traditional memory-augmented agents.

Existing code-generation benchmarks score a single mapping from a complete prompt to a one-shot output. However, real web development is different. Users seldom write a full spec at the start; many requirements only become clear once they look at an intermediate result and react to it. We present Asuka-Bench, a benchmark that pairs underspecified user intent with multi-round refinement, grounded in browser-rendered behavior. Each task is resolved through a closed loop: a Code Agent generates a web project, a UI Agent executes test cases on the deployed site, and a User LLM turns evaluation outcomes into natural-language feedback for the next round. The benchmark comprises 50 web tasks with 784 evaluation criteria and 2402 expected outcomes. We benchmark 8 LLMs across 2 agent frameworks. The results separate models clearly: weighted Task Pass Rate varies by 38 percentage points and models also differ substantially in their ability to repair from feedback. Asuka-Bench is also far from saturated: even the strongest model completes only 52% of projects after three rounds.

Self-evolving agents improve through continual self-play and self-generated learning signals, but autonomous evolution can also cause capability degradation and safety drift. Although human feedback has proven effective for static and post-trained agents, its role in self-evolving systems remains underexplored. We introduce Agent Norm Correction through Human-like Oversight and Review (ANCHOR), an LLM-based framework that simulates human supervision and delivers feedback at various phases of self-evolution. With ANCHOR, we evaluate two representative open-source self-evolving agent systems across coding, mathematical reasoning, and safety. Our results show that even limited supervision substantially mitigates safety degradation while preserving stable performance on core evolutionary objectives. Further analysis shows that supervision over the output verification phase is the most effective for intervention, whereas increasing supervision frequency yields diminishing returns. These findings provide empirical evidence and practical guidance for designing more stable, controllable, and human-aligned self-evolving agent systems.

Large language model agents increasingly rely on external tools, but larger tool menus can reduce reliability and efficiency by increasing wrong-tool calls, premature actions, and token cost. Existing tool-selection methods often optimize semantic relevance, exposing tools whose names or descriptions match the user request. We argue that relevance is insufficient: a tool may be related to the task while still being unnecessary or premature at the current step. We propose Causal Minimal Tool Filtering (CMTF), a training-free method that selects tools by causal sufficiency. CMTF uses lightweight precondition-effect contracts to expose only the minimal next-step tool frontier needed to advance from the current state toward the user goal. Across multi-step tool-use tasks, we compare CMTF with all-tools exposure, keyword retrieval, state-aware filtering, and causal-path ablations, measuring task success, wrong-tool calls, premature actions, tool exposure, and token cost. In the main benchmark with 102 tasks, 100 tools, four LLM backends, and 2448 task-method-model runs, CMTF matches the strongest causal baseline in aggregate success while reducing visible tools from 100 to one per step and reducing token usage by about 90% relative to all-tools exposure.

A central challenge for language agents is utilizing past experience to adapt to dynamic test-time conditions. While recent work demonstrates the promise of agentic memory mechanisms, most systems restrict retrieval to episode initiation. Consequently, agents are forced to rely on static guidance that becomes increasingly misaligned as long-horizon tasks unfold. To address this rigidity, we propose the Adaptive Memory Agent (AdaMEM), a novel framework for agent test-time adaptation. Without updating model parameters online, AdaMEM adapts agent behavior via a hybrid memory architecture: it maintains a long-term trajectory memory of raw experiences collected offline while generating dynamic short-term strategy memory on-the-fly to guide decision-making. This mechanism enables the trade-off between token efficiency and adaptability across varying inference-time compute levels. Empirically, AdaMEM significantly outperforms static memory baselines, achieving relative gains of up to 13% on ALFWorld and 11% on WebShop, with consistent leading performance extending to agentic search on HotpotQA. To further enhance this adaptation, we develop STEP-MFT, a Step-wise Memory Fine-Tuning technique that trains the policy to synthesize high-quality strategies from retrieved experiences, yielding additional performance gains. Our work establishes a new scaling dimension for agentic memory, supporting continuous reasoning and self-evolution post-deployment in real-world environments. Our code is available at https://github.com/yunx-z/AdaMEM.

Personal AI agents increasingly rely on long-term memory to provide persistent personalization across sessions. However, existing memory pipelines are largely driven by semantic similarity: memory data close to the current query is retrieved and injected into the model context. This creates a critical trustworthiness gap, since a semantically related memory may still be contextually inappropriate, leading to threats such as cross-domain leakage, sycophancy, tool-call drift, or memory-induced jailbreaks. In this paper, we study memory search as a trust boundary in personal AI agents. We evaluate representative agentic memory frameworks, including A-Mem, Mem0, and MemOS, together with OpenClaw, a real-world personal-agent environment with persistent state and tool-use capability. Our results show that long-term memory is not merely a utility layer, but a durable control channel that can reshape how agents interpret tasks and execute actions, leaving them highly susceptible to the aforementioned threats. To mitigate these vulnerabilities, we propose MemGate, a lightweight and deployable memory plug-in for trustworthy memory search, with only 9M parameters and a 35.1MB footprint. MemGate is inserted between the vector memory store and the backbone LLM, requiring no LLM modification, memory-database rewriting, or inference-time LLM judge. It applies a query-conditioned neural gate to candidate memory representations, turning raw similarity search into task-conditioned memory admission. Across multiple mainstream memory frameworks, real-world agent settings, and diverse LLM backbones, MemGate reduces memory-induced threats while preserving long-term memory utility.

Persistent memory for an LLM agent is a write-heavy substrate: every belief update is a versioned write, and a new claim may contradict a stored one. Production systems use four resolution heuristics (last-writer-wins, evidence-weighted merge, await-confirmation, per-rule policy), yet none declares the isolation level it assumes or the write-time anomalies it admits. We show that contradiction resolution is write-time concurrency control and make the missing contract explicit. TOKI types the four heuristics as one family of bitemporal operators over a dual-row schema, each with an isolation precondition and a provenance annotation that preserves the losing fact in an audit row. Four soundness theorems close the contract across isolation, schema, and provenance, lift the guarantees to operator pipelines, and extend the fold operators to n-ary conflict sets. A tightness companion proves that, within the relational schedule model, keyed logging of the adjudicating judge is necessary for replay consistency, which every audited baseline omits. A verdict matrix over eight systems localizes the gap: every baseline that keeps a language-model judge on the write path admits at least one of three write-time anomalies (replay inconsistency, belief-drift skew, audit erasure); a content-addressed engine-layer comparator avoids them only by removing the judge, and TOKI alone excludes all three while keeping it. On its one natural-workload slice the audit-row defence moves LoCoMo by 0.86, and ablating the typed memory layer removes 0.49 accuracy on 1,444 answerable LoCoMo questions; the cross-system comparison stays underpowered and claims no superiority. The contribution is the contract: a write-time correctness specification, proved sound across isolation, schema, and provenance, pinning the guarantee every production heuristic assumes but no deployed system makes explicit.

LLM-based agents increasingly rely on harnesses that provide execution environments, tool interfaces, context, lifecycle orchestration, observability, verification, and governance. Existing self-improving agents and automatic harness evolution methods mainly improve agents through runtime supervision, prompt optimization, workflow search, or harness modification based on final outcomes. However, they often fail to diagnose where the responsible evidence lies in failed trajectories and which harness layer causes the unreliable behavior, resulting in broad, indirect, or poorly scoped changes. This paper proposes HarnessFix, a trace-guided framework for diagnosing agent failures and repairing agent harnesses. HarnessFix compiles raw execution traces and harness code into a Harness-aware Trace Intermediate Representation (HTIR), which normalizes fragmented trajectory evidence and captures step-level provenance and control-flow relations. It then attributes failures to responsible trajectory steps and harness layers, consolidates recurring diagnoses into actionable flaw records, and maps them to scoped repair operators. Finally, HarnessFix generates and validates harness patches under flaw-specific repair specifications to reduce target flaws without introducing unacceptable regressions. We evaluate HarnessFix on SWE-Bench Verified, Terminal-Bench 2.0 Verified, GAIA and AppWorld. Across these benchmarks, HarnessFix improves held-out test performance over the initial harnesses by 15.2%--50.0%, outperforms human-designed and self-evolution baselines, and reveals recurring harness-flaw patterns across ETCLOVG layers.

Iterative retrieval-reasoning agents have recently shown promise for multimodal long-document question answering. However, most existing systems maintain a single growing context that mixes retrieval traces, observations, and intermediate reasoning. As interactions accumulate, key evidence becomes scattered and diluted, making multi-hop reasoning noisy. We propose MARDoc, a Memory-Aware Refinement Agent framework that decouples long-document QA into three specialized agents: an Explorer for multi-granularity multimodal retrieval, a Refiner for distilling interaction traces into structured evidence and reasoning memories, and a Reflector for checking evidence sufficiency and providing targeted feedback. Across iterations, the agents rely on a dynamically updated structured memory rather than a full accumulated interaction history. This design reduces context noise while preserving answer-critical facts and their logical dependencies. Experiments on MMLongBench-Doc and DocBench show that MARDoc achieves strong results, outperforming same-backbone baselines and demonstrating the effectiveness of structured memory for agentic document QA.

Despite recent progress, LLM agents still struggle with reasoning over long interaction histories. While current memory-augmented agents rely on a static retrieve-then-reason paradigm, this rigid pipeline design prevents them from dynamically adapting memory access to intermediate evidence discovered during inference. To bridge this gap, we propose MRAgent, a framework that combines an associative memory graph with an active reconstruction mechanism. We represent memory as a Cue-Tag-Content graph, where associative tags serve as semantic bridges connecting fine-grained cues to memory contents. Operating on this structure, our active reconstruction mechanism integrates LLM reasoning directly into memory access, allowing the agent to iteratively explore and prune retrieval paths based on accumulated evidence. This ensures that memory retrieval is dynamically adapted to the reasoning context while avoiding combinatorial explosion caused by unconstrained expansion. Experiments on the LoCoMo benchmark and LongMemEval benchmark demonstrate significant improvements over strong baselines (up to 23%), while substantially reducing token and runtime cost, highlighting the effectiveness of active and associative reconstruction for long-horizon memory reasoning.

LLM agents are increasingly deployed on long-horizon tasks requiring sustained reasoning over extended interaction histories. Realizing this at scale requires agents to persistently store, retrieve, and update their own memory across sessions. A rich ecosystem of agent memory systems has emerged spanning flat retrieval, LLM-mediated extraction, consolidating fact stores, and agentic control flows. Yet, their system-level behavior remains uncharacterized. We present the first systems characterization of agent memory. First, we introduce a system-oriented taxonomy classifying agent memory systems along four axes. Second, we build a phase-aware profiling harness attributing cost to construction, retrieval, and generation. Third, we characterize ten representative systems across two benchmark suites, uncovering how design choices shift cost across the write and read paths. Finally, we derive 10 system recommendations covering construction scheduling, capability floors, amortization via query volume, freshness-latency tradeoffs, and fleet-scale management.

Recent Large Language Models (LLMs) have shown impressive reasoning abilities; but they are still susceptible to hallucinations, intermediate reasoning mistakes, and unreliable reasoning results in complex mathematical reasoning problems. In this study, we introduce a critic-based heterogeneous multi-agent approach to improve the dependability of mathematical reasoning. This framework incorporates several LLM agents of different specialties and employs a critic-driven adaptive learning system to assess and guide the reasoning process based on intermediate feedback. The system adopts a generator-validator framework, with the validator not only determining correctness but also offering critiques to guide regeneration of solutions. This allows for adaptive error correction and prevents error cascading. Our experiments on the GSM8K benchmark show that the proposed method achieves up to 13% accuracy improvement over single-shot and non-critic models. Additionally, findings suggest that heterogeneity and critique reduce the need for large models, allowing smaller models to perform on par. Ablation studies reveal the main performance gains are due to the critic-based feedback loop and not model size. In summary, the proposed approach showcases the benefits of combining heterogeneous multi-agent collaboration and critique to obtain reliable and interpretable reasoning systems.

For over half a century, software engineering has operated on a foundational premise: human engineers decompose problems, encode decision logic into static code, and manually adapt that code as requirements evolve. This paper argues that the emergence of AI agents — systems where large language models serve as the primary reasoning engine, dynamically generating and discarding code as an instrumental resource — constitutes not an incremental improvement but a fundamental restructuring of the software paradigm. Drawing on first-principles analysis of complexity scaling, we formalize the distinction between traditional software (where code is the carrier of decision logic) and agentic systems (where code is ephemeral tooling for an LLM-driven reasoning loop). We trace the historical arc from licensed software to SaaS to what we term Agent-as-a-Service (AaaS), showing that each shift transferred additional complexity away from end-users. We introduce the concept of Agentic Engineering as an emergent discipline — distinct from software engineering in its core object of study, control model, and human role. Through analysis of recent benchmark evidence including SWE-bench Verified, EvoClaw, and LangChain's multi-agent coordination studies, we demonstrate both the transformative potential of the agentic paradigm and its current limitations. We conclude with a four-stage roadmap toward self-evolving agent ecosystems and concrete recommendations for practitioners navigating this transition.