AI Software Engineering

AI coding agents powered by LLMs are increasingly integrated into real-world software development, where they generate, edit, and execute code with autonomous access to local files and tools. Coding agents inherit security risks from both the LLM backbone, where adversarial prompts, poisoned training data, and backdoor triggers can cause models to emit insecure or attacker-chosen code, and their agentic architecture, where tool-using autonomy enables induced misuse of external APIs, data exfiltration, and persistent compromise of development environments. This paper presents a systematic evaluation of malicious issue requests against state-of-the-art coding agents (Cursor, Claude Code, and Codex Desktop), powered by two major model families (OpenAI GPT-5.3 Codex/GPT-5.4 and Anthropic Sonnet 4.6). Our novel benchmark IssueTrojanBench contains malicious issues that are constructed based on four novel attack categories (i.e., embedded as malicious instructions in issues), six delivery vectors (e.g., PDF, or issue comment), and further augmented by perturbations. Our results reveal critical vulnerabilities in the as-deployed modern coding agents, i.e., 66.5% of the malicious issues from IssueTrojanBench penetrate all the guardrails (agent- and LLM-level) of coding agents. Our further analysis shows that rejection is almost entirely from LLMs rather than the agent frameworks, with GPT models broadly vulnerable and Sonnet 4.6 exhibiting more selective, risk-aware blocking of high-impact actions. Our evaluation also highlights that the current agent-level defense strategy offers limited additional protection for coding agents. Our findings highlight the urgent need for stronger agent- and model-level safety mechanisms to protect AI coding agents.

Developers increasingly use two coding agents together: one writes a draft, and the other reviews it. However, it is not clear whether the pairing is worth its cost and time, or whether the order of the pairing matters. We run a controlled experiment on 116 recent hard and medium lcb tasks with Claude and Codex across six conditions to approximate a software practitioner's workflow: both solo baselines, both cross-model orderings, and both same-model orderings. The reviewer sees the problem and the writer's draft but cannot execute tests, which approximates a code review step. Claude review raises Codex drafts from 71.6% to 89.7% ($p_{BH}=.001$); Codex self review raises them to 84.5% ($p_{BH}=.022$). The reverse direction does not pay off: Codex reviewing Claude drafts drops the pass rate from 91.4% to 82.8% ($p_{BH}=.046$), and Claude self review leaves the 91.4% baseline unchanged. Our evaluation indicates that the useful pairing is asymmetric: use Claude to review Codex, not the other way around.

Small language models and coding agents increasingly generate web front-end code, yet their outputs are typically evaluated primarily for functional correctness. A generated interface may compile, render, and pass unit tests while still violating established interface quality principles, including accessibility barriers, deceptive design patterns, poor visual hierarchy, and excessive decision complexity. Existing auditing approaches face a trade-off between cost, coverage, and scalability: expert human review provides rich judgment but is slow and expensive; frontier vision-language models offer broader reasoning capabilities but remain costly to deploy at scale; and rule-based tools such as axe-core and Lighthouse are inexpensive but primarily capture mechanically checkable accessibility issues. We investigate whether a lightweight vision-language model can serve as an effective critic for generated interfaces. We unify 19 interface-quality principles from three complementary sources of HCI knowledge: WCAG 2.2 accessibility standards, deceptive design taxonomies, and established theories of perception, cognition, and interaction. To train this critic, we construct a verified dataset of approximately 10,000 generated web pages by synthetically injecting known violations into clean, LLM-generated Tailwind pages. Continued reinforcement learning on a 4B vision-language model improves micro-F1 from 36% to 84%, with 13 of 19 principles exceeding 80% F1. The resulting critic can audit generated interfaces, filter low-quality interface training data, and provide a reward signal for design-aware code generation. We release our data-generation recipe and injection/verification prompts to support reproducible evaluation and future work on scalable interface-quality assessment.

Large Language Models (LLMs) have demonstrated strong capabilities in code generation and reasoning, yet their ability to perform temporal reasoning over digital waveform data remains largely unexplored. Although reasoning over digital waveforms is a critical bottleneck in design verification, existing benchmarks primarily evaluate hardware description language (HDL) code generation and use waveforms only as supplementary context. This paper presents WaveformQA, an open-source question-answering benchmark for evaluating LLM temporal reasoning over digital waveforms. The benchmark comprises 360 questions with programmatically generated ground truths across eight categories of varying difficulty, including questions targeting multi-signal correlation and event ordering. Waveforms are generated from open-source design implementations, ensuring reproducibility and grounding the benchmark in real hardware behavior. Evaluation of frontier LLMs reveals that while models achieve reasonable accuracy on simple queries, performance degrades due to context window limitations and reasoning difficulties on complex temporal and multi-step questions. In addition, we show that an event-time JSON representation of waveforms improves LLM reasoning accuracy versus the standardized value change dump (VCD) format. The open-source framework supports extending to new question categories and importing new waveform sources, enabling researchers to rapidly prototype temporal reasoning experiments.

Traditional query processing engines require continuous development and extensions to support new techniques and user requirements, and in some cases, entirely new systems must be built from scratch. However, these engines are difficult to extend due to their internal complexity, and building new systems demands significant engineering effort and cost. To address this, we demonstrate GenDB, a generative query engine that shifts query processing from manually engineered systems to query processing code generation driven by Large Language Models (LLMs). An early prototype of GenDB uses LLM agents to generate instance-optimized query execution code tailored to specific data, workloads, and hardware resources. This prototype suits offline code generation for repetitive, templated queries, since the upfront generation cost amortizes over many executions and correctness can be ensured through extensive fuzz testing and manual inspection. For ad-hoc queries, GenDB can work with a traditional DBMS in a hybrid architecture: the DBMS handles one-off queries, while GenDB speeds up frequent SQL templates. Our demonstration allows users to (1) visually and interactively explore how GenDB analyzes workloads, profiles hardware resources and underlying data, produces query plans, generates code based on them, and finally uses an optimizer to iteratively achieve a correct and efficient implementation; (2) use visual inspection and analysis to gain qualitative insights into why GenDB produces code that achieves significantly better performance than state-of-the-art query engines on two benchmarks: TPC-H and a newly constructed benchmark designed to reduce potential data leakage from LLM training data; and (3) upload their own data and queries to explore GenDB with different LLMs and query patterns.

Current 3D generative models mostly produce a final surface: a visually strong but largely opaque mesh. Interactive 3D worlds need more than a surface. They need named parts, an assembly hierarchy, measurable constraints, local edit handles, and joints for articulation. We present Nova3D, a system that generates 3D assets as executable Blender source code; the compiled mesh, a binary glTF (GLB), is treated as the artifact, not the asset. Because the output is a program, semantic handles exist at generation time rather than being recovered afterward by segmentation or rigging. We evaluate on Nova3D-Bench, a frozen, spec-grounded benchmark of 54 items across six domains and three difficulty levels with text and image inputs, against eleven baselines in four families (mesh-native, part-structured, code-native, and CAD) plus a same-LLM ablation. Nova3D produces an executable program and a valid artifact for 54/54 items. Every asset exposes named parts organized in a parent-child assembly tree; no mesh-native, CAD, or segmentation baseline exposes either. It satisfies 51/52 prompt-stated numeric and count constraints (best baseline: 11/52), passes 14/18 blinded local edits with locality preserved in 18/18, and articulates 59 joints across 12 assets at 98.3% geometric validity, where every baseline exposes zero native joints. Its geometry is competitive: it wins the structured domains in a pairwise shape-quality tournament and is second only to the strongest mesh-native model, while conceding texture realism to baked-PBR systems. The central result is representational: code-native generation turns a generated 3D object from an opaque surface into a programmable asset that downstream systems can inspect, measure, edit, and animate.

Large language model (LLM) agents now perform well on correctness-oriented repository-level tasks, including SWE-Bench issue resolution and feature implementation in real codebases. However, they still struggle with repository-level code optimization, which requires preserving behavior while improving runtime performance. Passing tests is not enough in this setting; a patch must preserve behavior, implement code optimization, and approach expert speedups. Current agents often miss bottlenecks hidden behind abstraction layers and native extensions, stop after shallow speedups, or insufficiently test the code patches that thus may silently break edge cases. We present PerfAgent, a profiler-guided, verifier-in-the-loop workflow that gives an off-the-shelf coding agent the feedback needed to find real hotspots, improve beyond the first passing patch, and use profiler evidence rather than timing alone to decide what to optimize next. On two challenging optimization benchmarks, GSO and SWE-fficiency-Lite, PerfAgent more than doubles the rate of expert-matching patches over OpenHands with GPT-5.1, improving from 19.6% to 39.2% on GSO and from 26% to 74% on SWE-fficiency-Lite. It also surpasses an oracle best-of-five baseline at substantially lower cost, showing that the gains come from better feedback rather than additional test-time sampling.

Federated Learning (FL) enables collaborative model training without centralizing raw data, but building and operating FL systems remains difficult due to distributed execution, rapidly evolving frameworks, and privacy and governance requirements. In this paper, we present an empirical study of FL developer challenges by independently analyzing 495 Stack Overflow posts and 9,116 GitHub issues and pull requests from 92 FL-related projects. Using BERTopic-based topic modeling and difficulty indicators such as unresolved rates and median resolution time, we characterize recurring problem areas and compare how they manifest across the two support platforms, Stack Overflow and GitHub. Our analysis surfaces nine dominant Stack Overflow topics and thirteen GitHub topics, with persistent difficulties concentrated in environment setup and dependency compatibility, API breakages and migration, training instability under non-IID data, evaluation and metric correctness, and the integration of privacy-preserving mechanisms. We also categorize posts by question intent to understand the kinds of help developers seek; this intent analysis shows that "How"-type questions dominate, reflecting strong demand for procedural guidance. Several topics, such as "TFF Installation and Environment Compatibility" and "Federated Feature Engineering and SecureBoost Issues," exhibit high unresolved rates and long resolution times, suggesting shortcomings in tooling, documentation, and debugging support. Based on these findings, we provide actionable implications for FL framework designers, documentation authors, and educators. Although our results are constrained to public discussions and a subset of widely discussed frameworks, the study offers a scalable method for continuously monitoring developer pain points and improving the usability, reliability, and deployability of FL systems.

Coding agents increasingly operate in executable environments where a failed attempt produces actionable feedback rather than merely an incorrect answer. Existing cost-aware systems typically treat such failures as cascade decisions: try a cheap model first, then escalate hard cases to a stronger and more expensive model. In coding, however, execution feedback can also make further cheap-model recovery worthwhile, raising a budgeted deployment question: when should an agent spend more cheap compute, and when should it escalate? We formulate this post-failure decision as recovery routing over heterogeneous actions and train a supervised router from execution rollouts. To make the same router usable under changing budgets, we add a Conformal Risk Control (CRC) layer that selects a deployment-time cost penalty without retraining and provides marginal expected-cost control under exchangeability. Across held-out failures from five coding benchmarks, cheap recovery and escalation exhibit complementary success patterns. The calibrated frontier improves over fixed actions, prompt-only routers, and a binary cascade baseline; in the main GPT-5.4-nano/GPT-5.4 setting, one CRC-calibrated frontier point exceeds always-escalate solve rate while using 35% of its mean recovery cost. Code is available at https://github.com/Qijia-He/agent-budget-control.

Large language models (LLMs) excel at general-purpose code generation, yet how well they handle scientific code remains an open question. Existing datasets and benchmarks are limited in scale, domain coverage, or executable verification, leaving the true gap between current LLMs and reliable scientific code generators inadequately assessed. To address these limitations, we present SciCodePile, the largest scientific code corpus to date, constructed from 37,737 public repositories and collectively comprising 128GB of code that spans multiple computational science disciplines. From this corpus, we further curate an executable benchmark of 200 tasks, each equipped with a sandboxed execution environment and an automated test harness for functional verification. We evaluate 15 LLMs from both open-source and closed-source families on three tasks: prefix-to-suffix completion, fill-in-the-middle infilling, and executable code generation. Results show that scientific code generation remains highly challenging: The best CodeBLEU reaches only 38.13 and 38.37 on the two completion tasks, while the strongest model achieves just 12.30% Pass@1 on the executable benchmark, underscoring how far current models remain from reliable scientific code generation. To demonstrate the training utility of SciCodePile, we further show that continued pretraining on our corpus improves CodeBLEU by $\times$2.84 on scientific code completion, and instruction tuning on our data improves Pass@1 by $\times$4.79 on the executable benchmark. All code and data are available at https://huggingface.co/SciCodePile.

LLM coding agents operate by constructing trajectories that accumulate reasoning, tool calls, and results to enable multi-step decision-making. However, the conventional append-only trajectory architecture found in practice tightly couples file-read actions with their observations, capturing snapshots that become permanently fixed in the chronological history. As files change through agent edits or concurrent human modifications, these snapshots become stale, causing reasoning errors and causing agents to redundantly re-read files, with each re-read appending yet another copy to the trajectory. To mitigate this, we propose CORVUS, a novel trajectory architecture that decouples file-read actions from their observations by maintaining a synchronized registry of relevant files and injecting only their current contents at each reasoning cycle. This structural change produces significantly lighter-weight trajectories that remain synchronized with the actual codebase state by construction, eliminating redundant file copies and stale snapshots that bloat conventional trajectories. We evaluated CORVUS on SWE- POLYBENCH_VERIFIED and SWE-BENCH PRO across four LLMs, achieving 9-50% reduction in average input tokens per task, 15-32% shorter final prompts, and up to 37% fewer reasoning cycles while maintaining comparable pass rates.

Pruning long context for coding agents has been a vital technology for efficient context management. While existing context pruning methods such as SWE-Pruner realize this by attaching a separate code classifier, we find the agent itself encodes internal representations indicating the relevance of code context when reading tool output. Based on this finding, we propose SWE-Pruner Pro, which prunes tool outputs directly inside the agent. Concretely, a small head turns the agent's own internal representations into a keep-or-prune label for each line, with a length-aware embedding keyed to each tool output's line count. Across two open-weight backbones and four multi-turn benchmarks, SWE-Pruner Pro saves up to 39% of prompt and completion tokens while preserving task quality, with bounded inference overhead. Notably, on MiMo-V2-Flash SWE-Pruner Pro additionally raises the SWE-Bench Verified resolve rate by +3.8% and the long-context Oolong accuracy by +2.2 points.

Real-time multimodal applications, including voice agents and interactive video generation, compose heterogeneous models into pipelines whose efficient deployment requires application-specific decisions about placement, streaming, and intra-model parallelism. Existing serving systems and auto-parallelism compilers commit to limited transformations and fixed workload assumptions, so achieving high performance on a new application requires hand-crafting an efficient implementation. We present FlashRT, an agent harness that guides coding agents to lift simple developer-written reference implementations into optimized multi-GPU deployments that flexibly weigh target metrics like latency and throughput. Using a new chain-of-program paradigm, FlashRT directs a generic coding agent through a multi-pass transformation process where an agent transforms the reference into an intermediate representation (IR) to capture data dependencies and persistent-state scopes, validates this IR via a sequential interpreter, and performs static analyses to identify candidate transformations. Then, the agent iteratively implements, verifies, and benchmarks each candidate under a measurement-gated optimization loop to produce effective deployments that span different hardware budgets. Across various applications, including video world models and multimodal LLMs, FlashRT converts reference implementations into highly efficient deployments, delivering up to ~70x latency reduction and 2.8x throughput improvement on NVIDIA B200 GPUs. On AMD MI355X GPUs, FlashRT matches the peak latency reduction while increasing peak throughput improvement to 3.6x, demonstrating that agent-driven optimization can be more scalable on platforms with less mature expert optimization. In fact, for Qwen3-Omni text-to-audio inference, FlashRT reduces response latency by 65% compared to the expert vLLM-Omni implementation on AMD MI355X.

Coding agents are increasingly used to accelerate code generation in many downstream tasks, such as fixing bugs, building applications, and prototyping. However, despite their value as coding assistants, agent-generated code tends to be larger and more verbose than the corresponding human-written implementation. In this work, we show that the cause lies in the agent's own search process: while iterating toward a passing solution, an agent accumulates speculative edits, abandoned hypotheses, and temporary changes that persist into the final patch. This may seem harmless for a single patch, but the problem compounds as agents take responsibility for ever-larger portions of a codebase-a codebase that was once minimal and well-maintained slowly accumulates redundancy faster than it can be cleaned up, drifting to a state that is harder to maintain. Given the magnitude of this problem, we take a step towards alleviating this issue. First, we formally define this phenomenon as CodeSlop-the residual and functionally unnecessary edits commonly seen in AI-generated code. We then introduce our algorithm TRIM (Trajectory-guided Redundancy Identification and Minimization). Rather than minimizing CodeSlop directly, TRIM instead minimizes agent trajectories. As we show empirically, this indirect technique of minimizing CodeSlop is highly effective: TRIM cuts CodeSlop by 17.9%-32.9% across agentic scaffolds, with negligible performance regression. TRIM is also highly efficient, requiring roughly half the validation cost of algorithmic baselines such as Delta Debugging.

Coding agents can now be left alone to improve software against a score. In this pattern--recently popularized as "autoresearch"--the agent receives a dataset, an evaluation script, and one editable file, and iterates without supervision: modify the code, measure, keep the change if the score improves. But what does the agent actually optimize--the developer's intent, or the literal number? We ran this loop on a real production task: deciding which Quranic verses appear in a noisy speech-recognition transcript and splitting the transcript by verse. Two frontier coding agents, Claude Code and OpenAI Codex, started from the same blank file with the same instructions, budget, and reasoning effort, three runs each. Both independently invented the same algorithm (canonicalization, n-gram anchoring, dynamic-programming alignment)--and then diverged. Claude stopped early with compact, general code. Codex drove the score ~10x lower, largely by memorizing answers to individual evaluation rows (19-41 hardcoded verse ids per run): a clean natural instance of specification gaming by a production agent. In a preregistered second study, we added a held-out test set and told both agents it existed. The memorization vanished, and the score gap vanished with it--yet Codex's general core transferred better and more consistently (held-out detection+split 0.085+/-0.004 vs. 0.121+/-0.031), losing only on one missed rejection of non-recitation input. Two exploratory community arms (Cursor, Antigravity) are consistent with the pattern. Every agent's held-out solution matched or beat the hand-engineered pipeline it was built to replace--the best by an order of magnitude--and now runs in production. From the ways agents exploited our harness--reading sibling runs through shared git state, leaving notes to "future runs" in persistent memory--we distill five design rules for evaluating autonomous agents.

Maintaining up-to-date code documentation is difficult in fast-moving repositories because design knowledge is scattered across source files and pull requests. We present CODENS , a system that turns pull requests into living, accessible, and queryable documentation for production codebases. CODENS incrementally builds a typed software knowledge graph from pull requests, enriches components through schema-driven semantic extraction, derives typed relations between them, and exposes the resulting knowledge through three retrieval modes, including agent-guided graph traversal for repository-level question answering. The system also preserves semantic change history across pull requests and integrates both answer-quality and operational evaluation metrics. We evaluate CODENS on a client Ruby on Rails project in production. Results show that CODENS produces highly relevant and well-grounded answers, while qualitative feedback highlights a remaining challenge in concise, documentation-oriented synthesis.

Modern software teams have mature tools for low-level testing, such as pytest, JUnit, and Jest, which make it inexpensive to write unit tests and run them on every commit. Systems engineering, in parallel, has developed rigorous principles for design verification and validation (V&V), which has worked very well across engineering discipline to align user expecations and requirements with developers' deliverables. In practice, however, the two rarely connect, and the link between users' high-level requirements and the low-level tests that machines actually run is maintained by hand, if at all. This gap is increasingly costly for AI-enabled and cyber-physical systems, for which regulators now ask for traceable evidence that high-level requirements are met, while raw test results provide little of the structure such evidence requires. We introduce VNVSpec, an open-source framework that makes V&V specifications machine-readable and executable. With this framework, users state high-level requirements directly or import them from catalogs derived from published standards. Then, the framework checks requirement quality, supports decomposition into module-level requirements with explicit metrics and acceptance criteria, links these requirements to test results through a traceability graph, and compiles the collected evidence into verdicts and audit-ready reports. We evaluate the framework by self-application, in which it is continuously assessed in CI against its own specification of 36 requirements verified by 449 tests, completed within limited time which scales linearly and thus can handle up to 10,000 requirements. We also discuss how the framework extends to testing black-box AI models and AI coding agents. The framework, its full test suite, the catalogs, and the benchmark scripts are available at https://github.com/ai-vnv/vnvspec.

Developers frequently write uninformative git commit messages such as "fix" or "update stuff", degrading the value of version-control history for code review, debugging, and onboarding. We present CommitLLM, a three-stage pipeline that generates concise, Conventional Commits-compliant messages from code diffs using a fine-tuned small language model. The system combines (1) QLoRA fine-tuning of Mistral-7B-Instruct-v0.2 on the CommitPackFT dataset, (2) constrained decoding to enforce brevity, and (3) deterministic post-processing to strip conversational artifacts and enforce format. On a 50-sample evaluation, CommitLLM achieves 98% format compliance (vs. 22% for vanilla Mistral), reduces average output length from 154.8 to 37.9 characters, and improves LLM-as-a-Judge scores from 1.97 to 3.68 out of 5. Notably, the post-processing layers contribute more to quality improvement than the fine-tuning itself, suggesting that for structured-output tasks, treating the LLM as a component in a deterministic pipeline is more effective than optimizing the model alone. The entire system runs on a single consumer GPU (NVIDIA T4, 16 GB VRAM).

Biographical personas are widely used in system prompts, but their effects on code generation are rarely evaluated under controlled, pre-registered conditions. We tested four prompt conditions (no persona, two engineer personas, and a research-librarian persona), 12 code-generation tasks, two frontier models, and five runs per cell (480 completions). Persona effects differed between the two tested models. Under the pre-registered mixed-effects analysis, the condition-by-model interaction was significant for provider-reported output tokens; a post-hoc visible-character measure showed the same qualitative pattern. Six GPT-5.5 completions were length-capped and are reported separately. On Claude Opus, the minimalist engineer persona reduced visible output by 30% (33% in provider tokens) without improving correctness, while the thorough engineer persona increased output without a correctness gain. In an exploratory post-hoc analysis, the librarian persona elicited in-character disclaimers in 55 of 60 Opus responses and 12 genuine no-code responses, lowering mean correctness from 0.92 to 0.67. GPT-5.5 produced neither behavior in its 59 non-truncated responses. These results are consistent with personas acting as Model-Dependent behavioral-policy biases rather than universal quality interventions. We release raw completions, derived scores, analysis artifacts, a pre-registration document, and an execution gate log; end-to-end test-based rescoring requires an unreleased task harness.

When does a committed intermediate stage in an LLM reasoning pipeline earn its cost? Constrained Path Reasoning (CPR) pairs a source-aware path hypothesis with stage-level accounting. Search generates provisional states; trusted or validated invariants can constrain hard, while other proposals remain soft and revisable. CPR predicts that task-compatible commitments can factor transitions, concentrate candidate mass, induce regularity, and expose feedback when their gains exceed propagated error and execution cost. The formalism covers discrete commitments and continuous flows and measures effective branching, endpoint concentration, and cost per usable output. Across 1,180 generated QCQPs and 40 engineered degenerate polynomial instances (2,140 endpoints), residual triage recovers 63.0% of repair-all's additional feasible yield with 17.7% of its attempts. Fixed-LLM accounting (270 unique calls shared across nested arms) finds usable yield of 41.1% direct, 90.0% after formalization and deterministic execution, 20.0% after one-shot convexification, and 21.1% for the full path. In 120 paired-condition calls, a two-action rollback rule reaches 90% usable yield versus 36.7% for the feedback-conditioned selector. Two endpoint probes separate source from validation: a 72-output cross-trajectory transplant reduces entropy and acceptable mass; a 24-output same-call self-proposal pilot gives unchanged two-repeat collision entropy, 25.0% versus 8.3% usable yield, and 1/8 deterministically confirmed endpoint checks. Model-generated states supply hypotheses; trusted execution earns constraint strength.

Supervised fine-tuning (SFT) of open-weight LLMs on expert agent trajectories has emerged as a prominent approach to building capable code agents without reliance on proprietary models. A central yet underexplored question is how trajectory quality and quantity jointly shape model performance. We present a systematic empirical study of trajectory data filtering for LoRA fine-tuning of Qwen2.5-Coder-7B-Instruct on the SWE-trajectory dataset (67,074 trajectories, of which 32,161 are resolved). We propose a two-axis quality scoring framework -- Efficiency and Style -- and evaluate it through 16 controlled experiments spanning strategy, scale, and ablation analyses. Since 7B-scale models attain near-zero SWE-bench resolve rates, we adopt cross-entropy (CE) loss on held-out trajectories as the primary metric, validated via first-action generation: CE loss and ROUGE-L are perfectly rank-correlated (Spearman $ρ$ = -1.00), with limited-sample evidence supporting but not conclusively establishing this proxy. Our results reveal a scale-dependent quality-quantity trade-off: at small scales, doubling the dataset (500 to 1,000) yields ~12.7% CE-loss reduction whereas the TopQ-Random gap stays <1% (Mann-Whitney p > 0.10); at 2,000 trajectories this same gap widens to 3.6% (p = 0.016). Ablation further identifies error-retry rate as the dominant sub-dimension, performing comparably to the full composite ($Δ$ < 0.2%). Together, these findings establish trajectory-level quality scoring as a viable but scale-sensitive lever for code-agent SFT and offer a proxy-validated evaluation protocol for the regime where end-to-end resolve rate is statistically infeasible.

Serverless multi-model LLM systems multiplex popularity-skewed model catalogs over shared GPU pools, yet typically schedule each request independently. Tool-using agents break this abstraction: a session repeatedly calls an LLM across short tool gaps, carries a long reusable KV prefix, and is judged by session completion time (SCT). Load-only routing can separate a continuation from both its model and KV state, while round-based model multiplexing can delay even a correctly placed continuation until the target model's next slot. Both failures are especially costly for hundred-billion-parameter models: their weights constrain residency, while long-context KV is expensive to reconstruct or move. We present Talaria, a session-aware serverless multi-model serving system that makes session continuity a joint placement-and-admission decision. Its router ranks placements by model residency, KV locality, and instance pressure, while soft reservations account for likely returns in the last serving instance's admission budget. Session-prefill (SP) admits budget-eligible continuations before the active model slot closes. An instance-local substrate keeps HBM addresses stable, preserves host-restorable KV, and stages weights across model switches. On a single TP=8 server, we replay 30 SWE-Bench model-sessions (960 calls) over three models, each with more than 100B total parameters. Against an otherwise identical round scheduler with SP, host-KV restoration, and D2D staging disabled, Talaria cuts p50 SCT from 1000 s to 189 s and p95 from 2296 s to 867 s, speedups of 5.3x and 2.6x.

Code search enhances developer productivity by enabling efficient code reuse. Current code search systems often use a retrieve-then-rerank pipeline, where rerankers focus on modeling semantic relevance between queries and code. However, these rerankers overlook critical non-functional qualities like execution speed, memory usage, and maintainability, which are essential for practical software development. Studies reveal developers expect results to maintain high coding standards and satisfy specific needs, such as resource optimization, highlighting the importance of quality-aware code search. Achieving quality-aware code search faces two major challenges: the scarcity of quality-annotated datasets for effective training and the limitations of standard contrastive learning objectives, which fail to capture the ordinal relationships among high-quality, low-quality, and irrelevant code. Although contrastive learning excels in distinguishing relevant from irrelevant code, its binary objective does not support nuanced quality distinctions.To address these challenges, we propose SynH-Rank, a quality-aware code reranking framework that combines LLM-driven diverse data synthesis with hierarchical ranking training. SynH-Rank employs a three-level labeling scheme to explicitly model the hierarchy: high-quality relevant > low-quality relevant > irrelevant. Additionally, we introduce a new benchmark with 4,209 pairs and two novel metrics: Quality Preference Accuracy (QPA) for assessing prioritization of high-quality code and Multi-Condition Accuracy (MCA) for evaluating performance under complex constraints.Experimental results show SynH-Rank improves QPA by 20.15% over backbone models and outperforms standard relevance-only contrastive training by 15.80%, while simultaneously enhancing traditional relevance metrics and multi-condition generalizability.

Global LiDAR-to-BIM initialization must place a robot within an as-designed building model without a prior pose. In repetitive interiors, the principal failure mode is not low-confidence registration but confident aliasing: a submap can match several BIM regions with comparable scores, producing a high-scoring pose displaced by symmetry. We present Move First, Commit Later, a selective layer that treats a registration front-end as an evidence source and decides whether to commit. Candidates from multiple submaps are mapped to a common $\mathrm{SE}(2)$ anchor; a top-1 consensus $B_m$, invariant to non-champion multiplicity, aggregates cross-submap evidence; and topology serves only as a binary feasibility gate. The decision is typed--COMMIT, DEFER, or AMBIGUOUS($τ$), reporting the detected symmetry period--and reversible: symmetry-breaking motion upgrades AMBIGUOUS to COMMIT. On a real multi-room building and a controlled symmetric simulation, the layer commits correctly in every tested trajectory-scale configuration and otherwise abstains with a typed state, whereas forced-choice policies on the same front-end select wrong rooms in most cases. Committed anchors are within 0.02--0.36 m of an independent laser-tracker position reference and within $1.9^\circ$ of a BIM-registration orientation proxy. The evaluation covers one building and one front-end; the layer is designed to be front-end modular.

The increasing complexity of modern software systems has made automated code generation a fundamental task in software engineering. However, existing approaches often fail to adequately capture the intricate, multi-level dependencies among code entities, leading to generated code that is logically incomplete or difficult to integrate into real-world systems. To address this limitation, we propose a dependency-aware code generation framework that explicitly models interactions among code entities through a graph-based representation. We decompose dependencies into two complementary components: a quantized matrix that captures strong, explicit relations, and a sparse low-rank factorization that models weaker, implicit interactions. The decomposition is efficiently learned via an alternating optimization procedure. During code generation, the learned dependency structure is incorporated as a constraint, ensuring both semantic coherence and structural consistency of the generated code. Furthermore, we introduce a sparse triplet representation for strong dependencies, significantly improving storage efficiency and computational scalability. Extensive experiments demonstrate that our approach consistently produces code with superior semantic alignment and structural fidelity compared to existing methods.

Large language models (LLMs) are increasingly used to automate data-processing workflows, yet coding agents typically produce scripts that are not automatically materialized as persistent, editable platform artifacts. We call this disconnect the \textit{NL2Pipeline gap}. To bridge it, we introduce \textsc{DataFlow-Harness}, a platform that guides an LLM agent to construct platform-native directed acyclic graphs (DAGs) through typed, incremental mutations rather than free-form scripts. The platform combines \textsc{DataFlow-Skills} for procedural guidance, a Model Context Protocol (MCP) layer that exposes the live operator registry and current pipeline state, and \textsc{DataFlow-WebUI}, which synchronizes conversational authoring with a visual DAG editor. On a 12-task data-engineering benchmark, \textsc{DataFlow-Harness} achieves a 93.3% observed end-to-end pass rate. Relative to Vanilla Claude Code, it reduces measured monetary cost by 72.5% and generation latency by 49.9%; its observed pass rate is within 0.9 percentage points of the Context-Aware Claude Code baseline while its cost is 42.8% lower. Per-task analysis indicates that Skills are most useful when construction depends on implicit procedural knowledge. These results show that live platform grounding can produce persistent, editable workflow artifacts with an observed reliability close to script-generation baselines and with lower measured construction cost and latency.

Large-scale AI datacenter platforms comprise thousands of heterogeneous hardware components whose validation requires comprehensive fault injection test plans. Today these plans are authored manually: engineers review hardware self-healing validation documents and bills of materials, enumerate failure modes per field-replaceable unit, and produce flat lists of single-layer test cases. This process is labor-intensive, error-prone, and dependent on institutional knowledge; coverage gaps surface late, traceability to source specifications is implicit, and the effort is largely repeated per platform. This paper presents a generative AI multi-agent architecture that automates the generation of structured hardware validation test plans from two canonical inputs: self-healing validation documents, which enumerate known failure modes and their detection and remediation behaviors per field-replaceable unit, and component Bills of Material. An ingestion agent normalizes heterogeneous inputs into a canonical representation; a classification agent maps components to functional domains via contextual reasoning over part descriptions and sub-category hierarchies; and a generation agent synthesizes test cases by combining normalized failure modes with domain-classified data, filling gaps and producing edge cases. The output conforms to a standardized schema for direct import into internal validation software. Evaluated on two production platforms against manual baselines, the framework achieves coverage expansions of 74.2% and 51.4%, cutting authoring from days to hours. It yields fully traceable mappings from each test case to its source specification, and its multi-agent decomposition is portable across platform generations. Automated and expert evaluations confirm 100% extraction fidelity and high acceptance of new scenarios, validating the framework as a robust human-in-the-loop force multiplier.

Moderation systems increasingly rely on external toxicity tools, but those tools are unreliable under code-mixing, transliteration, slang, and language mismatch. We study the \emph{conditional reliability} of toxicity priors in Indian multilingual and code-mixed short text: English toxicity, Indic abuse, and rule-based severity cues can be useful evidence, but only in some linguistic and abuse-severity contexts. We propose ToxGate, a trust-fusion head that conditions each auxiliary signal on the encoder representation before adding it to the prediction state. Across three short-text abuse datasets, four transformer encoders, and five seeds per setting, ToxGate improves over matched plain encoders in 10 of 12 in-domain settings and 7 of 8 transfer settings. The largest and most interpretable gains occur in high-risk moderation slices, including explicit slurs, violent threats, and cross-dataset transfer. The broader lesson is that moderation systems should treat external toxicity tools and priors as conditional evidence rather than fixed features or ground truth, in focused ablations, source-specific gating gives the strongest results in transfer, severe-abuse slices, and high-risk triage.

Coding agents can fix a failing example without preserving the domain rule that made it fail. We present agentic synthesis against counterexample-supplemented sketches, a repository-native method for systems whose policy is discovered during implementation. A human starts with a partial sketch, and a coding agent compiles a replaceable projection. When simulation exposes missing or mistaken policy, an operator approves the corrected behavior and the minimum general rule the case authorizes. Every Developer call names its change authority and the rules, holes, anchors, and approved behavior that must survive. Conflict or ambiguous permission leaves the files unchanged and produces a clarification question. A complete archive preserves provenance; a curated regression set gates distinct boundaries. Before another candidate is revealed, the active case and curated regressions must pass both deterministic approved-output comparison and a separate review against the current sketch. Periodic clean regeneration tests whether the sketch carries the learned policy. We demonstrate the method with CatSynth, a captured synthetic application. In one open-world run with GPT-5.4-mini, 8 of 14 frozen candidates became counterexamples. Under the corrected protocol, replay-all, evolved-sketch rebuild, and retained Sketch-CE each passed all 8 accepted cases. They passed 14, 17, and 16 of 21 withheld cases, respectively. Sketch review rejected premature empty-input and tag policies and restored dropped anchors; adjudicated reviewer errors did not become policy. One model and one reveal order cannot establish general correctness or superiority. On this suite, the second check exposed drift hidden by deterministic replay, and the reviewed sketch passed three more withheld cases than raw example replay.

Large Language Models (LLMs) perform differently on identical programming tasks when prompted in different natural languages, a phenomenon known as language bias. While this behavior has been widely studied for general text generation, its impact on code generation quality and programming conventions remains largely unexplored. We investigate how the language used to describe programming tasks affects the source code generated by GPT-4o mini, DeepSeek, and Claude. Our study comprises 460 coding tasks spanning Python (230) and Java (230). We translate and manually curate the original English prompts into Chinese, Hindi, Spanish, and Italian while preserving their technical meaning. We evaluate the generated code using multiple dimensions, including functional correctness through test pass rates, structural quality using established code metrics, issues detected by static analysis tools, and lexical characteristics such as the language used in identifiers and comments. Our results show that (i) English prompts do not consistently produce the best functional correctness or code quality, (ii) the impact of prompt language depends on both the programming language and the LLM, and (iii) generated code frequently mixes English with the prompt language in comments and string literals. These findings provide the first curated multilingual benchmark for studying language bias in code generation and offer insights for developing more robust multilingual code generation systems.

Large language models (LLMs) exhibit a well-documented gap between latent capability and consistent activation: the router hypothesis posits that models possess the knowledge to solve a task but lack reliable internal routing to activate it. Prior work in formal mathematical reasoning (SAIR, Cázares 2026) reports that structural priors (cheatsheets) raise in-distribution performance dramatically, yet collapse below the zero-shot baseline out-of-distribution (OOD) -- and that iterative recalibration amplifies rather than corrects the collapse. We test whether this phenomenon is cross-domain by reproducing the SAIR design in source-code security vulnerability detection, evaluating three LLMs (GPT-OSS-120B, Llama-3.3-70B, Gemma-4-31B) across three vulnerability categories (CWE-798, CWE-284, and the non-CWE N+1 anti-pattern) spanning syntactic, contextual, and semantic complexity, then transferring cheatsheet-augmented prompts to real-world CVE data from VUDENC (CWE-89, CWE-22). Our findings replicate and extend SAIR: (F1) structural priors lift semantic-vulnerability recall from 20.0% to 100.0% across all models; (F2) zero-shot performance degrades along a semantic complexity gradient; (F3) the same cheatsheets that saturate synthetic performance amplify distribution-shift collapse on real CVE data (CWE-89: 100% synthetic F1 to 48.9% on VUDENC, -51.1pp); (F5) iterative recalibration produces a v2 cheatsheet that performs worse than v1 on real data, mirroring SAIR's AN45c-vs-AN38 finding. These results provide evidence that the cross-distribution trade-off surface documented in SAIR generalises to code security, and that the router hypothesis is cross-domain. We argue the structural nature of the collapse motivates distribution-aware training over prompt calibration. Code and evaluation scripts: https://github.com/bytepro-ai/bitcoder-v2-research

Payment integration is a demanding repository-level software task: agents must select a suitable product, implement coordinated client-server flows, verify payment outcomes, and preserve consistency between transaction and business states. We introduce Alipay-PIBench, a benchmark for evaluating coding agents on realistic Alipay payment integration. It contains nine product-specific projects and 18 task instances, each organized into Basic functional-completion and Advanced risk-aware hardening scenarios. Scenario-specific rubrics support deterministic static, unit, integration, and end-to-end checks, supplemented by LLM-assisted assessment for semantic requirements. We evaluate six coding-agent models and report rubric pass rate (RPR). Under the with-skill condition, mean RPR ranges from 68.58% to 91.37%. Access to the alipay-payment-integration skill improves mean RPR by 10.31 percentage points on average relative to the without-skill condition, with gains varying across models, products, and scenarios. Method-level results distinguish source-level completion, executable payment behavior, and payment-domain requirements. Alipay-PIBench provides a controlled setting for diagnosing model capability and evaluating structured guidance in payment integration.

Coding agents now produce a growing share of a team's code, while the reasoning behind each change -- the alternatives weighed, the constraints discovered, the approaches rejected -- is trapped in assistant transcripts that vanish with the session. Memory for this setting, the agentic development lifecycle (ADLC), is usually posed as one retrieval problem and built as machinery: tiered stores, memory graphs, compiled wikis, model-judged admission. We argue memory should instead be git-bound -- built into the repository's version control, inheriting the guarantees the machinery struggles to construct: ground truth from commits, freshness from rebuild, verification from the merge, containment from review. On this ledger we solve two problems separately, then combine them. Seed supply is closed as an eight-corpus retrieval study under a pre-registered ship discipline: five imported ranking mechanisms rejected, two kept, and a best configuration of ~0.31 pooled MRR -- ~60x the raw-transcript grep floor, ~15x an honest parsed-turn floor. Answer assembly is where ranking stops helping: single-shot retrieval scores only 0.07-0.20 answer-sufficiency on real developer questions, and ungated episode injection measurably degrades good answers. A router dispatches breadth to a git-anchored structural map, pointed lookups to confidence-gated episodes, and rationale to decision synthesis, which reconstructs why-arcs no single session contains (0.83 sufficiency on a young ~50k-LOC production system). Routed, the system answers at 382-980 tokens per question -- three orders of magnitude below the recorded history. Because ground truth is mined from commit-session links rather than annotated, every result is replicable on any user's own history at zero labeling cost. The remaining constraint is capture. Code, benchmark, and paper source: github.com/rekal-dev/rekal-cli.

AI coding agents produce code faster than humans can review it. In our approach, the prover is the judge of whether the code is correct. Under a verifier-driven loop, AI agents wrote and verified bare-metal security software in Ada/SPARK spanning classical and post-quantum cryptography, TLS 1.3, IKEv2, X.509, and a Matrix client. GNATprove discharged 49,280 proof obligations, established functional correctness for selected primitives, and proved the absence of run-time errors for the rest, at roughly 20-40 times lower supervision cost than comparable hand verification. GNATprove alone was insufficient: some defects could not be detected and were resolved using known-answer tests, interoperability, or human review of specifications. Given weak checks, the agent tried to bypass them and reported success. We report where each layer caught faults and draw the central lesson: what an agent can be trusted to establish is bounded by the strength of its feedback.

Agentic coding tools are increasingly capable of generating and submitting pull requests (PRs) to software projects, introducing new forms of human-agent collaboration in software development. While prior studies have examined PR-level outcomes of agent-generated contributions, less is known about how agentic coding tools are adopted and managed at the project level. In this paper, we analyze 25,264 agentic PRs from 2,361 popular GitHub repositories to investigate (1) the adoption of agentic coding tools, (2) project-level agentic PR productivity, and (3) human-agent collaboration patterns. Our results show that the median repository generates only one to two agentic PRs during a three-month period, indicating that intensive adoption remains concentrated in a small subset of projects. At the same time, small projects (1-5 contributors) exhibit higher participation ratios and average levels of agentic PR activity than medium-sized and large projects. We also observe substantial variation in project-level agentic PR productivity. While a small number of projects exceed an industry-reported estimate of 36 PRs per participant during the three-month observation period, most projects remain below this threshold. Finally, human-agent collaboration is dominated by a single-human oversight model, in which one developer reviews and/or modifies the agent's contributions, while multi-human collaboration patterns remain uncommon. These findings provide early empirical evidence on how open-source projects organize human oversight around agentic coding tools and suggest that successful integration of agent-generated contributions depends not only on advances in agent capabilities but also on the human and organizational processes that govern their use. Because this study captures an early snapshot of agent adoption, future work should continue to track how adoption patterns evolve over time.

Languages with rich static semantics, such as Rust, provide stronger guarantees for AI-generated code, but their strictness makes generation more difficult. Off-the-shelf compilers can provide useful feedback post-generation, but does not guide intermediate generation steps, such as those during autoregressive LLM decoding. Constrained decoding intervenes earlier by rejecting invalid tokens during sampling, but requires white-box model access and costly reimplementation for semantic constraints.We introduce generative compilation, the first approach to obtaining compiler feedback on partial programs during generation. The core technical device is a sealor: a lightweight, mostly syntax-guided transformation that converts partial programs into complete ones that standard compilers can diagnose. It is designed such that possible-to-complete partial programs are never rejected, while preserving enough code context to catch genuine dead ends early. We construct such a sealor on a core Rust-like calculus and prove that it satisfies these properties, all mechanized in Lean. We extend it to the first partial-program checker for real Rust. We evaluate our method on challenging repository-level Rust coding tasks, across both frontier black-box and open-weight models. We show that generative compilation reduces non-compiling outputs and improves functional correctness, relative to standard post-generation feedback. It does so by detecting a broad range of errors close to their source and early during generation, thereby reducing errors cascades and enabling focused diagnostics. More broadly, generative compilation is a step toward making compilers a first-class citizen of AI-assisted programming active during generation, rather than a separate post-generation check.

The growing adoption of local inference frameworks such as Ollama has made it increasingly common for developers to run large code models on laptops and other resource-constrained hardware. In these settings, post-training quantization is essential for reducing memory footprint and enabling practical deployment, yet its impact on generated code remains insufficiently understood. We empirically evaluate six state-of-the-art quantization methods (GPTQ, AWQ, QuIP#, AQLM, BitsAndBytes, and GGUF) on two representative large code model families, Qwen2.5-Coder and CodeLlama, using the multilingual McEval and CoderEval benchmarks for Python and Java. We assess functional correctness (pass@1) together with maintainability, reliability, security, and structural complexity. We also introduce a novel analysis of robustness under varying prompt complexity, characterized by Shannon entropy and token length. Our results show that quantization techniques differ meaningfully in their impact on correctness and code quality. AQLM consistently matches or exceeds the full-precision baseline, whereas QuIP# exhibits the largest correctness degradation, particularly on complex prompts. Security attributes remain stable across models, benchmarks, and programming languages, while robustness to prompt complexity varies across techniques. These findings provide practical guidance for selecting quantization strategies for deploying large code models on resource-constrained hardware and highlight the importance of evaluating quantized models beyond functional correctness.

Formal contracts are essential for software testing and verification, yet writing them remains labor-intensive and error-prone. LLMs offer a promising path toward autoformalization: synthesizing executable assertions from natural-language specifications and thereby bridging the gap between informal developer intent and formal executable specifications. We present Monty: an autoformalization framework for assertions that tackles the challenges of expectations of validity of assertions and ambiguity in natural-language. Our techniques are based on filtering formalizations using a novel conformance score metric and validity scores obtained from testing the code against formalized assertions. We evaluate our approach on 541 assertion-generation tasks derived from 22 collection-like Java classes, and show that our technique produces the ground truth more reliably (improving upto 20 points in precision on average) than when using LLMs naively to translate assertions.

Distinguishing semantic-preserving commits from changing ones remains an open challenge in software repository mining. While existing approaches detect refactoring commits accurately, they cannot ensure that a commit is purely semantic-preserving, without any interleaving behaviour-changing modification. This limitation can impact several tasks, such as debugging, fault localisation, bug dataset construction, rollback analysis, and bug fixes backporting. To fill this gap, we propose SemaDiff, a novel approach for identifying semantic-preserving commits through behaviour-based analysis; comparison of similar test execution on pre- and post-commit versions. As code impacted by the refactoring is often hard to test and different accross both versions, we propose generating additional calling methods to that code, which serve as testing target. Given a commit, SemaDiff analyses the diff to identify modified code and extracts unchanged dependent code that calls it. It then generates an additional dependent class using a large language model to exercise the changed code in both versions, and automatically generates tests for the dependent code. This way, we obtain the same tests for the different code versions, enabling the behavioural-difference detection. The commit is classified as semantic-preserving only if all generated tests produce identical outcomes across the two versions. To evaluate SemaDiff, we construct and annotate manually a dataset of 183 commits, gathered from well-known open-source Java projects. The obtained results show that SemaDiff distinguishes accurately semantic-preserving from -- changing commits in about 76% of the cases, with a 100% precision in semantic-changing commit detection.

Coding agents must integrate external tool returns into ongoing reasoning - a capability that standard left-to-right pretraining on code exposes only in its forward direction. We observe that the action-observation-continuation loop of a coding agent is structurally isomorphic to a function call site, where a caller binds arguments, a callee returns a value computed elsewhere, and downstream code consumes that value. This conditioning structure exists at internet scale in ordinary code. We exploit it through function-aware fill-in-the-middle (FIM) mid-training: a self-supervised objective that masks functions selected via program dependency graph analysis and a complexity-inferability double criterion. We mid-train Qwen2.5-Coder-Instruct (7B/14B) and Qwen3-8B on a 2.6B-token decontaminated corpus drawn from 968 GitHub repositories, then apply existing agentic post-training pipelines. Mid-training improves SWE-Bench-Verified by +2.8/+3.0 at 7B/14B and by +3.2 on Qwen3-8B; SWE-Bench-Lite gains are +3.7/+4.0/+5.4 on the same models. The improvement holds across two post-training pipelines (R2E-Gym, SWE-Smith) and on a non-Qwen2.5 base (Qwen3-8B with SWE-Lego). Beyond in-domain gains, mid-training also mitigates the capability erosion that agentic post-training otherwise inflicts on non-agent coding (e.g., LiveCodeBench) and non-coding tool-use benchmarks (tau-bench, BFCL): although the mid-training corpus contains Python code only, the function-call inductive bias survives post-training and yields consistent gains.

Agent benchmarks often compare two agents after all tasks have run, but costly evaluations make partial runs tempting. A task fraction alone does not show whether a partial run supports the same pairwise conclusion as the completed benchmark. We study this question by replaying completed public task-level records from SWE-bench, AppWorld, and tau-bench. A partial budget counts as enough only when it supports the completed benchmark's decision, covers required task groups, and leaves no more than a target fraction of comparisons unresolved. The required task fraction varies sharply. At the strict 0 percentage point threshold on a 5 percentage point budget grid, AppWorld first meets all targets at 15 percent, tau-bench at 25 percent, and SWE-bench Verified at 90 percent; SWE-bench Lite does not meet all targets by 95 percent under the primary coverage rule. Partial-evaluation reports should state how much one agent must outperform another, how tasks are selected, what coverage rule is required, what decision rule is used, and how many comparisons may remain unresolved.

The rise of Artificial Intelligence (AI) enables automatic analysis of large amounts of data. Previously time-consuming and labor-intensive tasks can be completed much more efficiently with the use of AI. This work uses AI techniques to analyze and revise curricular patterns in an undergraduate degree for Software Engineering. Curricula often have long sequences where failure to pass a class within the sequence may jeopardize completion of the degree within four years. Manual analysis and revision of curricula by university faculty is a lengthy and labor-intensive process, causing changes to occur rarely and making it impossible to keep up with the changing needs of students. This work reduces the time-to-change for curricula and reduces bottlenecks and graduation delays by using Large Language Models (LLMs) to analyze curricular patterns and suggest revisions.

Synthesizing deployable robot behavior trees (BTs) from natural language (NL) requires grounding to ensure every generated BT references only skills a robot can actually execute. Existing LLM-based BT synthesis approaches often place this grounding responsibility on the prompt author. This makes deployment brittle when the author does not know which skills the robot can execute, how those skills are parameterized, or how the robot runtime software constrains valid BT structure. This paper proposes a contract-grounded BT synthesis architecture in which a coding agent queries a robot-side Model Context Protocol (MCP) server to retrieve an explicit contract consisting of a skill library, permitted BT operators, and optional BT composition templates, before synthesizing a BT for validation and execution. In our framework, non-expert operators issue NL commands without knowledge of robot implementation details, while a robot runtime validation gate enforces correctness before execution. We evaluate two LLMs, a closed model (Sonnet 4.6) and a smaller open-source model (Gemma4:31b), across 110 simulated tasks in PyRoboSim and 14 tasks on a physical Husarion Panther robot. Results show that contract grounding enables near-perfect BT validation and high task success, that BT composition templates substantially recover success on reactive control-flow tasks for the smaller model, and that the architecture transfers to physical hardware running a Nav2 stack opaque to both operator and agent.

LLM-based coding agents repeat the same classes of mistakes across sessions because they lack a mechanism to retain corrections from human review feedback. We present a closed-loop framework in which every accepted review comment is codified as a persistent behavioral rule, progressively expanding the set of error classes the agent can self-detect. The framework combines an accumulating rule set in a version-controlled instruction file, a self-review checklist executed before code submission, and automated validation that ensures rule set integrity as it grows. In deployment across a 35+ service microservices platform, the rule set grew from 5 to 18 behavioral rules, 15+ language-specific standards, and a 15-item self-review checklist, all derived from real review feedback. We present empirical results from 11 recorded working sessions spanning code generation, PR review, incident investigation, and cross service refactoring. We observe that accumulated rules shift review effort from low-level correctness toward design-level validation, achieve a measured 0% recurrence rate for ruled-against error classes, and transfer across heterogeneous agent interfaces. We compare our approach against related work in experiential LLM learning (Reflexion, ExpeL, Voyager) and automated code review (CodeReviewer, SWE-bench agents), showing that our framework achieves persistent cross-session learning without weight updates, operates on production codebases rather than synthetic benchmarks, and addresses an orthogonal dimension (behavioral consistency over time) that existing benchmarks do not measure. The result is a coding agent that improves with every review cycle, accumulating the engineering wisdom of its human collaborators without changing a single model weight.

Context-reduction layers for API-based coding agents, including command-output compressors, retrieval rankers, and API-boundary proxies, are commonly evaluated by how much context or tool output they remove. We ask a different question: which interventions actually reduce end-to-end billed cost while preserving task success? Our primary evidence is a pre-specified, hash-frozen, paired campaign of 2,908 provider-billed Claude Code runs, of which 2,848 were analyzed, covering 103 tasks, seven repositories, and three models. The campaign compared a baseline with two generations of hook-based compression and an API-boundary proxy within a broader measured program of roughly 5,500 billed executions. Three findings emerge. First, prompt-cache traffic dominated cost composition, accounting for about 87% of reconstructed four-component cost (about 80% of the actual bill), with an 8.7% dollar-weighted residual not attributable from retained telemetry. Second, local payload reduction was not a reliable predictor of end-to-end billed cost. An arm that removed 38% of estimated raw tool-output tokens incurred 6.8% higher paired cost (95% CI: +2.8% to +11.3%), while per-task reduction showed only a weak association with cost change (Pearson r = 0.15). Third, aggressive compression can remove action-critical evidence: on SWE-bench-derived Go tasks, compression reduced successful patch application from 27/40 to 15/40 by corrupting verbatim edit anchors. We propose evaluating context-reduction systems by success-adjusted billed cost rather than token reduction alone.

Can large language models perform technical comprehension of computer architecture papers--not summarization, but structured critique that names the core mechanism, surfaces buried assumptions, and connects a contribution beyond its own scope? We study Gauntlet, an open-source pipeline that analyzes a paper through five independent expert-persona reviewers and an adversarial synthesis stage. On 20 ISCA 2025 and HPCA 2026 papers, 10 researchers each wrote their own analyses and then judged, for papers other than their own, the human analysis against Gauntlet's. Across the 20 comparisons evaluators preferred Gauntlet in 15 (human in 4, one tie); its advantage is significant on per-analyst totals (two-sided Wilcoxon, p < 0.001) and largest on Critical Rigor. Where humans win, it is on trust and usefulness rather than depth: a confident wrong claim, a mechanism described but not taught, or unprioritized breadth. A 98-paper automated ablation shows the gain comes from the multi-agent structure: the pipeline beats the same model run as a single rich-persona agent on 96% of papers. We release all analyses, scores, and the rubric as a community resource.

Active flow control involves nonlinear dynamics, partial observations, and computationally expensive simulations, making controller design particularly challenging. Deep reinforcement learning (DRL) has emerged as a powerful framework for such problems, but its success typically relies on large numbers of simulator interactions and produces neural-network policies whose decision process often remains difficult to interpret. In this work, we investigate a different paradigm: instead of optimizing neural-network parameters, we use modern coding agents to search directly for explicit executable feedback laws. We introduce a constrained heuristic-learning protocol in which an agent iteratively proposes, evaluates, and revises controller implementations while interacting exclusively through the public benchmark interface. The proposed framework is evaluated on 13 active flow-control benchmarks spanning one, two, and three-dimensional problems and compared against the strongest available DRL baselines under identical simulation budgets. The discovered heuristic controllers match or outperform the best DRL policy in 10 of the 13 environments while remaining compact, interpretable, and directly inspectable. Beyond aggregate performance, the resulting controllers reveal physically meaningful feedback mechanisms, transfer successfully across more challenging configurations, and remain competitive under varying Reynolds and Rayleigh numbers, actuator counts, and observation sparsity. These results suggest that heuristic learning through coding agents constitutes a credible and complementary alternative to conventional reinforcement learning, combining competitive performance with physically interpretable controller representations. Prompts and source code are available at https://github.com/DonsetPG/fluid-heuristic-learning.

Recent autonomous penetration testing papers report high benchmark scores while adding multi-component security harnesses around frontier LLMs. Because these systems often change both architecture and backbone model, it is difficult to tell how much performance comes from the harness rather than from the underlying model. This paper presents a controlled study on the 104-task XBOW benchmark using default coding CLI agents as plain-agent baselines. We first run Codex, OpenCode, and Pi with the same GPT-5 model, budget, target interface, and scoring rule. This phase identifies the strongest same-model baseline and tests whether security-specific prompt variants improve its observed score. We then compare the default Codex scaffold with published MAPTA and PentestGPT V2 results under the closest available model matches. Finally, we repeat the plain-agent experiment with GPT-5.2 and GPT-5.5 to measure model scaling inside the same scaffold. The results show a mixed but practical picture. Specialised harnesses can add measurable benchmark lift and may improve cost efficiency, but plain coding agents already solve a large share of the benchmark; repeated plain-agent runs can match or exceed some published architecture scores in union coverage, and newer models substantially improve the same scaffold. Future evaluations should report model-matched plain-agent baselines before attributing benchmark gains to architecture design alone.

Despite their central role in fault detection, test oracles remain challenging to construct effectively. Recent learning based methods address this challenge by automatically generating test assertions, yet even if syntactically correct, they are often ineffective in revealing bugs. Rather than generating assertions, this study explores a different approach by training a model to directly predict whether a given test prefix passes or fails. We present FOCAL, an emerging code LLM-based discriminative oracle predictor. It learns from labeled pairs of test prefixes and methods under test, employs losses that emphasize failing cases during training, and grounds its predictions in statement level behavioral evidence. Compared with the baseline method SEER, we substantially improve performance on failing cases for unseen projects and provide richer explanations. A preliminary evaluation on fault-detection benchmarks and automated test-generation artifacts shows that our approach is highly accurate within its training distribution and substantially improves failure detection on previously unseen projects where prior discriminative oracles collapse. Moreover, the highlighted statements are supported by behavioral explanation checks. These early results suggest that fail-aware discriminative oracle prediction can complement existing approaches such as fuzzing, search-based testing, and LLM-based test generation. These techniques produce test prefixes at scale but often lack fault oriented oracles. In future work, FOCAL could take generated test prefixes and attach fault-aware predicted oracles to them, turning high-volume input generation into executable tests that are more likely to expose semantic failures.

Autonomous coding agents force engineering organizations to choose between API-based frontier models -- strong reasoning at high token cost -- and on-premise quantized open-weights models, which promise low-marginal-cost scaling and data sovereignty at some loss of reasoning fidelity. We study this trade-off through a single-developer, non-randomized longitudinal case study over two contiguous 28-day periods on a production monorepo: an API-based Claude Opus 4.7/4.8 configuration using Claude Code versus an on-premise GLM-5.1/5.2 configuration using Opencode, quantized to NVFP4, on NVIDIA Blackwell hardware. Analyzing LLM telemetry and Git history, we find that prompt caching (99.3% hit rate) cuts realized API cost by 88.6% to an effective $0.57 per million tokens -- below even the $2.83 amortized unit cost of the shared on-premise slice (a utilization-dependent inversion; total realized spend and total cost of ownership (TCO) are the robust quantities). At comparable gross code churn, the local configuration was associated with a far higher defect-repair burden: a Fix Commit Ratio (FCR) of 74.9% versus 45.9%, with the odds of a commit being a repair 2.6 to 4.9 times higher within every difficulty tier (Mantel-Haenszel OR = 3.61). Under Taiwan-market parameters and a symmetric labor model, on-premise deployment nonetheless saves 40.1% of true TCO under shared GPU allocation, whereas dedicated reservation costs 43.8% more than the cached API. Under shared allocation, the genuine penalty is not monetary but a measurable developer-experience burden -- timestamp indicators show more work trapped in debugging spirals and a slower commit cadence -- and an offline replay shows hybrid routing gateways trade defect rate for infrastructure savings along a cost-quality frontier rather than dominate the pure-API baseline.

Large language models (LLMs) are increasingly used in agentic coding settings, where they can inspect files, execute commands, run tests, observe failures, and iteratively revise code. This shift raises a central evaluation question: can an agentic LLM generate an end-to-end software artifact that is both deployable and behaviorally correct under execution? Backend services provide a controlled but realistic substrate for this evaluation. Their APIs expose application-level executable semantics, and deployed behavior can be checked deterministically against an OpenAPI contract through black-box HTTP interactions. We introduce BackendForge, a benchmark of 56 contract-defined backend generation tasks rewritten from real open-source applications. Given a visible specification and an OpenAPI contract, an LLM must generate a Dockerized service that is built, deployed, and evaluated only through HTTP tests. To strengthen evaluation without introducing hidden requirements, BackendForge uses a test agent and a code agent to co-evolve the test oracle and reference service, where the test agent proposes specification-grounded backend tests and the code agent repairs the reference implementation. Although the best-performing model, GPT-5.5, succeeds on 55.4% of tasks under the base oracle, it succeeds on only 28.6% under the final oracle. This gap suggests that current LLMs can implement many local API behaviors, but still struggle to produce complete backend services.

Large language models (LLMs) often produce homogeneous outputs, raising concerns that AI coding assistants may lead to convergence in the software artifacts that developers create. Whether this occurs in practice is unclear because developers interactively prompt, evaluate, modify, and reject model outputs, and because outputs vary with prompt and repository context. I examine code homogenization using Kaggle contest submissions from 2019 to mid-2026. I first document widespread convergence toward the random seed value 42, consistent with LLMs reinforcing a longstanding convention in programming culture. I then study homogenization more broadly, at two levels of aggregation and abstraction. At the submission level, I measure the average pairwise similarity of submissions within contests. At the contest level, I measure the conceptual span of submitted code, motivating distinct measures for each: TF-IDF representations, which capture surface syntax, and Voyage 3 code embeddings, which capture code intent and semantics. The results demonstrate substantial syntactic homogenization at both the individual and collective levels: individual submissions have become more alike in literal syntax and code structure, while the latent dimensionality of syntactic variation has narrowed. In contrast, I find little evidence of semantic homogenization, individually and collectively. Average semantic distance remains essentially flat, and the contest-level latent dimensional span of semantic approaches remains stable, with evidence suggesting it has even expanded modestly. These findings suggest that AI coding assistants are certainly standardizing implementation details, yet they have not yet produced evidence of homogenization in the approaches and problem-solving strategies coders employ.

The rise of Software Engineering (SE) agents, i.e., LLM-based agents that can understand large codebases and carry out engineering tasks with limited human intervention, has been marked by rapid advances and adoption, but little is known about how developers build these systems in practice: existing studies mine repositories or examine deployment, but few investigate how SE agents are constructed. Through semi-structured interviews with 20 practitioners from 12 organizations and an online survey of 80 practitioners, this paper is the first to study how SE processes are changing in the development of SE agents and what challenges developers face. We find that as implementation becomes cheaper, bottlenecks shift rather than disappear: long-standing non-coding work such as requirements, coordination, review, and deployment becomes more visible, while reviewing and evaluating agent output becomes new and central. We characterize a seven-stage workflow and a shift toward evaluation-driven development, in which evaluation steers iteration and specifications become versioned artifacts read by both humans and agents. We further identify six challenges that teams face, together with the practices they adopt to address them, including unreliable evaluation signals, comprehension debt as code outpaces understanding, and behavioral changes introduced by provider-side model updates.

Computational imaging, which recovers hidden signals from indirect, noisy measurements, underpins quantitative discovery across scientific disciplines, yet building a correct reconstruction pipeline demands deep domain expertise and remains laborious even for domain scientists. We introduce Imaging-101, a benchmark of 57 expert-verified computational imaging tasks spanning six scientific domains, each grounded in a peer-reviewed paper and canonicalized into a standardized four-stage pipeline (preprocessing, forward physics modeling, inverse solver, and visualization) Three evaluation tracks (planning, function-level unit tests, and end-to-end reconstruction) probe distinct agent capabilities across the full pipeline. Evaluating seven frontier LLMs uncovers systematic challenges in applying coding agents to computational imaging that go beyond those exposed by general coding benchmarks, spanning algorithm selection, physical convention handling, and pipeline integration. These findings highlight concrete capability gaps and point toward skill-augmented, domain-specialized agents as a practical path to reliable computational imaging assistance.

As AI code tools become integrated into programming environments, students increasingly describe intended behavior in natural language and rely on these tools to generate code, shifting emphasis from code writing to specification. Yet little is known about the comments students write as specifications in AI-assisted programming tasks. We analyze a four-year dataset of undergraduate programming submissions and reflections from tasks in which students wrote comments to guide code generation and refined solutions using test-case feedback. We introduce a taxonomy spanning three dimensions: comment type, code expression level, and code construct. Using automated classification, we examine how these dimensions vary across attempts and how students describe the process in their reflections. Our findings show that students mostly wrote natural-language What comments, shifted toward How comments for more procedural constructs, and focused more on verifying generated code than on repeatedly rewriting comments.

Modern coding agents expose multiple tool surfaces -- IDE primitives, bash, and Model Context Protocol (MCP) code-execution -- and the field has shipped three contradictory claims about which one matters. We run the missing crossed comparison: an integrity-clean three-arm ablation (baseline / bash_only / code_only) on synthetic computation tasks and SWE-bench Mini modification tasks, holding model, harness, and prompts fixed, with two agents (Claude Code, OpenAI Codex CLI) so the comparison spans both regime and agent-design axes. Across the four resulting (regime, agent) cells, restricting the agent to a single execute_code MCP tool is cheaper than -- or statistically tied with -- its cheapest tool-rich rival in three cells (significantly on Artifact/Claude and SWE-bench/Codex; directionally on Artifact/Codex), with pass rates statistically tied within each cell. The lone exception is SWE-bench/Claude, where code_only is directionally costlier (+14.4%, not significant); a conditional-cost analysis localizes that gap to failure-cost on doomed-run trajectories, not a per-edit tax on successful runs. Two implications: the cheapest tool surface is jointly determined by task regime and agent design rather than by either axis alone, and the headline cost signal lives in cache-adjusted cost -- not pass rate, which is invariant across surfaces at the model sizes we evaluate. The benchmark harness, task suite, and analysis code are available at https://github.com/hyang0129/onlycodes.

Multimodal BrowseComp tasks require agents to combine perception, tool use, and long-horizon reasoning over dynamic web content, challenging their ability to handle compositional structure, open-world uncertainty, and multimodal integration across extended interactions. Crucially, real-world multimodal browsing involves three distinct information-flow patterns: text-only, image-to-text, and text-to-image, yet existing data construction methods cover only the text-only and image-to-text patterns, leaving text-to-image largely unaddressed and limiting agent generality and robustness. We introduce UNIBROWSE, a unified data pipeline that for the first time simultaneously generates training data covering all three patterns, augments curated knowledge graphs with live web retrieval for improved fidelity, and introduces a novel metric of exploration degree to filter low-signal instances for efficient reinforcement learning. Through this pipeline, we produce high-quality cold-start tool-use trajectories and exploration-rich QA pairs, and train a 35B-scale agent via supervised fine-tuning and exploration-aware RL.The resulting UNIBROWSE agent achieves state-of-the-art performance on multimodal BrowseComp benchmarks, attaining an average accuracy of 54.4 across five diverse benchmarks -- an improvement of 10.5 points over its base model Qwen3.5-35B-A3B -- and surpassing serveral closed-source agent workflows such as GPT-5 (42.9), Gemini-2.5 Pro (44.8), and Gemini-2.5 Flash (41.3).

Large Language Models (LLMs) are increasingly used for code smell detection tasks due to their ability to interpret program semantics. However, their reliability in this context remains poorly explored, particularly under varying prompt conditions where model predictions may be influenced by external cues rather than code characteristics. One such limitation is sycophancy bias, where models tend to align their outputs with user-provided assumptions instead of performing objective analysis. In this paper, we present the first systematic empirical study of sycophancy bias in LLM-based code smell detection. Using the MLCQ dataset, we evaluate how different prompt framings like confirmation bias, contradictory hints, and false premises affect model predictions. Our results show that LLMs are highly sensitive to prompt variations, with Decision Flip Rates reaching up to 72% and False Alignment Rates exceeding 90%, indicating substantial instability and agreement with misleading prompts. To address this issue, we propose Evidence-Guided Debiasing Prompting (EGDP), a structured prompting strategy that enforces evidence-first reasoning. EGDP reduces decision instability and improves robustness, lowering Decision Flip Rates to as low as 12% and False Alignment Rates to as low as 21%, while increasing reliance on structurally grounded evidence. Our findings demonstrate that sycophancy bias poses a critical threat to the reliability of LLM-based code smell detection, and that evidence-guided reasoning provides an effective and generalizable mitigation approach.

With the rapid development of Large Language Models (LLMs), text watermarking has emerged as a crucial technique for identifying machine-generated content. However, directly applying existing logits-based watermarking methods to code generation remains challenging, since the low-entropy nature of code exacerbates the trade-off between code quality and watermark detectability. In this paper, we propose a novel code watermarking approach called Grammar-Driven Watermark (GDW) for LLMs. GDW preserves syntactic validity through a grammar-guided three-level masking mechanism and injects watermark signals via structural role-aware modulation, assigning a stronger bias to content-bearing tokens while applying a more conservative bias to syntax-critical tokens. Aligning with the generation process, we further design a role-aware weighted detection statistic to improve detectability. Experiments across multiple programming languages, models, and decoding strategies show that GDW establishes a stronger quality-detectability trade-off frontier than existing methods, while maintaining robustness against variable-renaming attacks.

The rapid emergence of the Model Context Protocol (MCP) has introduced a new standard for connecting large language models to external tools and services. Despite its rapid adoption in open-source development, systematic understanding of how MCP is implemented, structured, and maintained remains limited. This study presents the first large-scale, evidence-based dataset of real-world MCP implementation collected directly from GitHub. Using a hybrid pipeline that integrates the GitHub REST and GraphQL APIs with custom Python verification scripts, 3,238 candidate repositories were discovered, filtered, and validated through multi-stage evidence checks. Each verified project was classified by operational role (e.g., client, server, gateway) and exported in a reproducible JSONL schema. A manual review of a representative subset confirmed an overall precision of 83% at a 95% confidence level, and additionally revealed a set of repositories functioning primarily as educational samples, tutorials, or demonstration templates. A targeted exclusion rule was then applied to remove these non-operational repositories, resulting in a final dataset of 2,297 validated MCP projects. The analysis shows that Python and TypeScript dominate MCP development, with hybrid architectures emerging as the most common design pattern. By emphasizing transparent verification strategies, structured evidence tagging, and reproducible data organization, this work establishes a foundational benchmark for studying real-world MCP ecosystems and supports future research on integration, connectivity, and compatibility across the broader developer community.