{"@context":"https://schema.org","@type":"NewsArticle","generatedAt":"2026-07-23T07:21:26.498Z","headline":"AgentDebugX：面向LLM智能体的开源故障调试框架","description":"AgentDebugX是一个开源调试框架，将LLM智能体调试组织为\"检测-归因-恢复-重跑\"闭环。其核心DeepDebug在Who&When基准上对qwen3.5-9b达到精确的智能体与步骤归因准确率，在GAIA上单次重跑即可修复失败任务。该工具提供Python库、CLI、Web控制台和可安装的智能体技能。","url":"https://www.aioga.com/news/cmrvnxfx7047nbihbzr6crygg/","mainEntityOfPage":"https://www.aioga.com/news/cmrvnxfx7047nbihbzr6crygg/","datePublished":"2026-07-21T00:00:00.000Z","dateModified":"2026-07-21T00:00:00.000Z","inLanguage":"zh-CN","publisher":{"@type":"NewsMediaOrganization","name":"Aioga","url":"https://www.aioga.com"},"citation":["https://arxiv.org/abs/2607.18754","https://aihot.virxact.com/items/cmrvnxfx7047nbihbzr6crygg"],"canonicalUrl":"https://www.aioga.com/news/cmrvnxfx7047nbihbzr6crygg/","directAnswer":{"@type":"Answer","text":"来源摘要将AgentDebugX描述为面向LLM智能体的开源调试框架，采用“检测、归因、恢复、重跑”闭环，并称其提供Python库、CLI、Web控制台及可安装的智能体技能。","url":"https://www.aioga.com/news/cmrvnxfx7047nbihbzr6crygg/","dateCreated":"2026-07-21T00:00:00.000Z","author":{"@type":"Organization","@id":"https://www.aioga.com/authors/aioga-editorial/#editorial-team","name":"Aioga Editorial Team","url":"https://www.aioga.com/authors/aioga-editorial/"}},"evidence":[{"@type":"CreativeWork","name":"arXiv source article","url":"https://arxiv.org/abs/2607.18754","datePublished":"2026-07-21T00:00:00.000Z","provider":{"@type":"Organization","name":"arXiv","url":"https://arxiv.org/abs/2607.18754"}},{"@type":"CreativeWork","name":"AIHot archive record","url":"https://aihot.virxact.com/items/cmrvnxfx7047nbihbzr6crygg","datePublished":"2026-07-21T00:00:00.000Z","provider":{"@type":"Organization","name":"AIHot","url":"https://aihot.virxact.com/items/cmrvnxfx7047nbihbzr6crygg"}}],"aggregationSource":"HuggingFace Daily Papers（社区热门论文）","originalPublisher":{"name":"arXiv","url":"https://arxiv.org/abs/2607.18754"},"article":{"id":"cmrvnxfx7047nbihbzr6crygg","slug":"cmrvnxfx7047nbihbzr6crygg","url":"https://www.aioga.com/news/cmrvnxfx7047nbihbzr6crygg/","title":"AgentDebugX：面向LLM智能体的开源故障调试框架","title_en":"AgentDebugX： An Open-Source Toolkit for Failure Observability， Attribution， and Recovery in LLM Agents","summary":"AgentDebugX是一个开源调试框架，将LLM智能体调试组织为\"检测-归因-恢复-重跑\"闭环。其核心DeepDebug在Who&When基准上对qwen3.5-9b达到精确的智能体与步骤归因准确率，在GAIA上单次重跑即可修复失败任务。该工具提供Python库、CLI、Web控制台和可安装的智能体技能。","source":"HuggingFace Daily Papers（社区热门论文）","sourceUrl":"https://arxiv.org/abs/2607.18754","aiHotUrl":"https://aihot.virxact.com/items/cmrvnxfx7047nbihbzr6crygg","publishedAt":"2026-07-21T00:00:00.000Z","category":"产品更新","score":71,"selected":true,"articleBody":["arXivLabs is a framework that allows collaborators to develop and share new arXiv features directly on our website.","Have an idea for a project that will add value for arXiv's community? Learn more about arXivLabs ：https://info.arxiv.org/labs/index.html."],"articleImages":[{"sourceUrl":"https://arxiv.org/static/base/1.0.1/images/funders/simons-foundation.png","alt":"Simons Foundation","afterParagraph":1,"url":"/media/articles/cmrvnxfx7047nbihbzr6crygg/e2d7f38d62f5ca91.png"},{"sourceUrl":"https://arxiv.org/static/base/1.0.1/images/funders/simons-foundation-international.png","alt":"Simons Foundation International","afterParagraph":1,"url":"/media/articles/cmrvnxfx7047nbihbzr6crygg/1d56e29c5557cbdc.png"}],"mediaStatus":"ok","articleBodyZh":["arXivLabs 是一个框架，允许合作者直接在我们的网站上开发和分享新的 arXiv 功能。","有一个可以为 arXiv 社区增加价值的项目想法吗？了解更多关于 arXivLabs 的信息：https://info.arxiv.org/labs/index.html。"],"translationStatus":"translated","bodyOrigin":"source-page","editorial":{"summary":"来源摘要将AgentDebugX描述为面向LLM智能体的开源调试框架，采用“检测、归因、恢复、重跑”闭环，并称其提供Python库、CLI、Web控制台及可安装的智能体技能。","background":"标题与摘要聚焦AgentDebugX及其调试能力，但正文摘录实际介绍的是arXivLabs协作框架，未出现AgentDebugX、DeepDebug、Who&When、GAIA或相关实验细节。","viewpoint":"Aioga判断，现有材料存在摘要与正文摘录不对应的问题。摘要中的模型、基准、归因准确率和单次重跑修复等表述，尚缺少正文证据支持，不宜据此扩展结论。","implications":"若摘要信息准确，该框架可能为智能体故障定位和失败任务恢复提供统一工作流；但在论文正文、代码仓库或实验数据得到核验前，其实际效果与适用范围仍无法确认。","nextStep":"建议补充并核对论文摘要、方法章节、实验表格和开源仓库，确认DeepDebug、qwen3.5-9b、Who&When及GAIA相关表述，再决定是否发布性能结论。","evidenceRefs":["title","summary","articleBody","source"],"status":"published","aiGenerated":true,"autoApproved":true,"generatedBy":"aioga-editorial:gpt-5.6-sol","reviewedBy":"aioga-editorial-review:gpt-5.6-sol","generatedAt":"2026-07-22T23:49:13.746Z","sourceHash":"360eaf216078c46a","review":{"approved":true,"groundedness":96,"clarity":92,"duplicationRisk":18,"blockingIssues":[],"notes":["候选内容准确区分了来源摘要中的信息与正文摘录实际提供的证据，并对未经论文正文或实验数据核验的性能结论使用了审慎、条件性的表述。","“Aioga判断”中的“Aioga”缺少上下文说明，但属于可选的表述清晰度问题，不构成事实性阻断。"]},"validation":{"passed":true,"mode":"ai-auto","revisions":0,"checks":["schema","length","source-attribution","low-source-overlap","no-html","independent-ai-review"]}},"tags":["产品更新","HuggingFace Daily Papers（社区热门论文）"],"translations":{"zh-CN":{"title":"AgentDebugX：面向LLM智能体的开源故障调试框架","summary":"AgentDebugX是一个开源调试框架，将LLM智能体调试组织为\"检测-归因-恢复-重跑\"闭环。其核心DeepDebug在Who&When基准上对qwen3.5-9b达到精确的智能体与步骤归因准确率，在GAIA上单次重跑即可修复失败任务。该工具提供Python库、CLI、Web控制台和可安装的智能体技能。","category":"产品更新","source":"arXiv","aggregationSource":"HuggingFace Daily Papers（社区热门论文）","pageTitle":"AgentDebugX：面向LLM智能体的开源故障调试框架 - Aioga AI资讯","description":"AgentDebugX是一个开源调试框架，将LLM智能体调试组织为\"检测-归因-恢复-重跑\"闭环。其核心DeepDebug在Who&When基准上对qwen3.5-9b达到精确的智能体与步骤归因准确率，在GAIA上单次重跑即可修复失败任务。该工具提供Python库、CLI、Web控制台和可安装的智能体技能。","url":"https://www.aioga.com/news/cmrvnxfx7047nbihbzr6crygg/"},"en":{"title":"AgentDebugX: An Open-Source Fault Debugging Framework for LLM Agents","summary":"AgentDebugX is an open-source debugging framework that organizes LLM agent debugging into a \"Detect-Attribute-Recover-Rerun\" closed loop. Its core, DeepDebug, achieves precise agent and step attribution accuracy on the Who&When benchmark for qwen3.5-9b, and can fix failed tasks on GAIA with a single rerun. The tool provides a Python library, CLI, web console, and installable agent skills.","category":"Products","source":"HuggingFace Daily Papers（社区热门论文）","aggregationSource":"HuggingFace Daily Papers（社区热门论文）","pageTitle":"AgentDebugX: An Open-Source Fault Debugging Framework for LLM Agents - Aioga AI News","description":"AgentDebugX is an open-source debugging framework that organizes LLM agent debugging into a \"Detect-Attribute-Recover-Rerun\" closed loop. Its core, DeepDebug, achieves precise agen...","url":"https://www.aioga.com/en/news/cmrvnxfx7047nbihbzr6crygg/","contentTranslated":true,"sourceHash":"3316b9dd99dbe3c6","translatedAt":"2026-07-22T17:06:28.098Z"},"ja":{"title":"AgentDebugX：LLMエージェント向けのオープンソースデバッグフレームワーク","summary":"AgentDebugXはオープンソースのデバッグフレームワークで、LLMエージェントのデバッグを「検出-原因特定-復旧-再実行」のサイクルとして整理します。その核心であるDeepDebugは、Who&Whenベンチマーク上でqwen3.5-9bに対してエージェントとステップの正確な原因特定精度を達成し、GAIAでは単回の再実行で失敗したタスクを修復できます。このツールはPythonライブラリ、CLI、Webコンソール、そしてインストール可能なエージェントスキルを提供します。","category":"製品更新","source":"HuggingFace Daily Papers（社区热门论文）","aggregationSource":"HuggingFace Daily Papers（社区热门论文）","pageTitle":"AgentDebugX：LLMエージェント向けのオープンソースデバッグフレームワーク - Aioga AIニュース","description":"AgentDebugXはオープンソースのデバッグフレームワークで、LLMエージェントのデバッグを「検出-原因特定-復旧-再実行」のサイクルとして整理します。その核心であるDeepDebugは、Who&Whenベンチマーク上でqwen3.5-9bに対してエージェントとステップの正確な原因特定精度を達成し、GAIAでは単回の再実行で失敗したタスクを修復できます。...","url":"https://www.aioga.com/ja/news/cmrvnxfx7047nbihbzr6crygg/","contentTranslated":true,"sourceHash":"3316b9dd99dbe3c6","translatedAt":"2026-07-22T17:06:44.868Z"},"ko":{"title":"AgentDebugX: LLM 에이전트를 위한 오픈소스 오류 디버깅 프레임워크","summary":"AgentDebugX는 오픈소스 디버깅 프레임워크로, LLM 에이전트 디버깅을 '감지-원인 분석-복구-재실행'의 폐쇄 루프로 조직합니다. 그 핵심인 DeepDebug는 Who&When 벤치마크에서 qwen3.5-9b에 대해 에이전트와 단계 귀속 정확도를 달성했으며, GAIA에서는 단일 재실행으로 실패한 작업을 복구할 수 있습니다. 이 도구는 Python 라이브러리, CLI, 웹 콘솔 및 설치 가능한 에이전트 기능을 제공합니다.","category":"제품 업데이트","source":"HuggingFace Daily Papers（社区热门论文）","aggregationSource":"HuggingFace Daily Papers（社区热门论文）","pageTitle":"AgentDebugX: LLM 에이전트를 위한 오픈소스 오류 디버깅 프레임워크 - Aioga AI 뉴스","description":"AgentDebugX는 오픈소스 디버깅 프레임워크로, LLM 에이전트 디버깅을 '감지-원인 분석-복구-재실행'의 폐쇄 루프로 조직합니다. 그 핵심인 DeepDebug는 Who&When 벤치마크에서 qwen3.5-9b에 대해 에이전트와 단계 귀속 정확도를 달성했으며, GAIA에서는 단일 재실행으로 실패한 작업을 복구할 수...","url":"https://www.aioga.com/ko/news/cmrvnxfx7047nbihbzr6crygg/","contentTranslated":true,"sourceHash":"3316b9dd99dbe3c6","translatedAt":"2026-07-22T17:07:34.912Z"},"es":{"title":"AgentDebugX: Marco de depuración de fallos de código abierto para agentes LLM","summary":"AgentDebugX es un marco de depuración de código abierto que organiza la depuración de agentes LLM en un ciclo cerrado de \"detección-atribución-recuperación-reintento\". Su núcleo DeepDebug alcanza una precisión de atribución exacta de agentes y pasos en qwen3.5-9b sobre el benchmark Who&When, y en GAIA basta un solo reintento para corregir tareas fallidas. Esta herramienta ofrece biblioteca de Python, CLI, consola web y habilidades de agente instalables.","category":"Productos","source":"HuggingFace Daily Papers（社区热门论文）","aggregationSource":"HuggingFace Daily Papers（社区热门论文）","pageTitle":"AgentDebugX: Marco de depuración de fallos de código abierto para agentes LLM - Aioga Noticias de IA","description":"AgentDebugX es un marco de depuración de código abierto que organiza la depuración de agentes LLM en un ciclo cerrado de \"detección-atribución-recuperación-reintento\". Su núcleo De...","url":"https://www.aioga.com/es/news/cmrvnxfx7047nbihbzr6crygg/","contentTranslated":true,"sourceHash":"3316b9dd99dbe3c6","translatedAt":"2026-07-22T17:07:32.983Z"},"fr":{"title":"AgentDebugX : cadre open source de débogage des pannes pour les agents LLM","summary":"AgentDebugX est un cadre de débogage open source qui organise le débogage des agents LLM en une boucle fermée \"détecter-attribuer-récupérer-relancer\". Son module central DeepDebug atteint, sur le benchmark Who&When, une précision de l'attribution des agents et des étapes pour qwen3.5-9b, et sur GAIA, une seule relance suffit pour corriger une tâche échouée. Cet outil fournit une bibliothèque Python, une interface en ligne de commande, une console Web et des compétences d'agent installables.","category":"Produits","source":"HuggingFace Daily Papers（社区热门论文）","aggregationSource":"HuggingFace Daily Papers（社区热门论文）","pageTitle":"AgentDebugX : cadre open source de débogage des pannes pour les agents LLM - Aioga Actualités IA","description":"AgentDebugX est un cadre de débogage open source qui organise le débogage des agents LLM en une boucle fermée \"détecter-attribuer-récupérer-relancer\". Son module central DeepDebug...","url":"https://www.aioga.com/fr/news/cmrvnxfx7047nbihbzr6crygg/","contentTranslated":true,"sourceHash":"3316b9dd99dbe3c6","translatedAt":"2026-07-22T17:08:17.475Z"},"de":{"title":"AgentDebugX: Open-Source-Fehlerbehebungsframework für LLM-Agenten","summary":"AgentDebugX ist ein Open-Source-Debugging-Framework, das das Debugging von LLM-Agenten als einen geschlossenen Kreislauf von 'Erkennen-Zuschreiben-Wiederherstellen-Neu ausführen' organisiert. Sein Kern DeepDebug erreicht auf dem Who&When-Benchmark eine präzise Zuschreibungsgenauigkeit für Agenten und Schritte bei qwen3.5-9b und kann auf GAIA fehlgeschlagene Aufgaben mit einem einzigen Neustart beheben. Das Tool bietet eine Python-Bibliothek, CLI, Webkonsole und installierbare Agentenfähigkeiten.","category":"产品更新","source":"HuggingFace Daily Papers（社区热门论文）","aggregationSource":"HuggingFace Daily Papers（社区热门论文）","pageTitle":"AgentDebugX: Open-Source-Fehlerbehebungsframework für LLM-Agenten - Aioga KI-News","description":"AgentDebugX ist ein Open-Source-Debugging-Framework, das das Debugging von LLM-Agenten als einen geschlossenen Kreislauf von 'Erkennen-Zuschreiben-Wiederherstellen-Neu ausführen' o...","url":"https://www.aioga.com/de/news/cmrvnxfx7047nbihbzr6crygg/","contentTranslated":true,"sourceHash":"3316b9dd99dbe3c6","translatedAt":"2026-07-22T17:08:21.028Z"},"pt-BR":{"title":"AgentDebugX: Um framework de depuração de falhas de código aberto voltado para agentes LLM","summary":"AgentDebugX é um framework de depuração de código aberto que organiza a depuração de agentes LLM em um ciclo fechado de 'detecção-atribuição-recuperação-reexecução'. Seu núcleo, DeepDebug, atinge uma precisão de atribuição de agentes e etapas no benchmark Who&When para qwen3.5-9b, e no GAIA, uma única reexecução é suficiente para corrigir tarefas falhas. Esta ferramenta fornece uma biblioteca Python, CLI, console web e habilidades de agente instaláveis.","category":"产品更新","source":"HuggingFace Daily Papers（社区热门论文）","aggregationSource":"HuggingFace Daily Papers（社区热门论文）","pageTitle":"AgentDebugX: Um framework de depuração de falhas de código aberto voltado para agentes LLM - Aioga Notícias de IA","description":"AgentDebugX é um framework de depuração de código aberto que organiza a depuração de agentes LLM em um ciclo fechado de 'detecção-atribuição-recuperação-reexecução'. Seu núcleo, De...","url":"https://www.aioga.com/pt-BR/news/cmrvnxfx7047nbihbzr6crygg/","contentTranslated":true,"sourceHash":"3316b9dd99dbe3c6","translatedAt":"2026-07-22T17:09:04.384Z"},"ru":{"title":"AgentDebugX: открытая платформа для отладки ошибок для интеллектуальных агентов LLM","summary":"AgentDebugX — это открытая отладочная платформа, которая организует отладку LLM-агентов в замкнутый цикл «обнаружение — атрибуция — восстановление — повторный запуск». Ее ядро DeepDebug на бенчмарках Who&When достигает точной атрибуции агентов и шагов для qwen3.5-9b, а на GAIA повторный запуск один раз позволяет исправить неудачные задачи. Этот инструмент предоставляет библиотеку Python, CLI, веб-консоль и устанавливаемые навыки агентов.","category":"产品更新","source":"HuggingFace Daily Papers（社区热门论文）","aggregationSource":"HuggingFace Daily Papers（社区热门论文）","pageTitle":"AgentDebugX: открытая платформа для отладки ошибок для интеллектуальных агентов LLM - Aioga Новости ИИ","description":"AgentDebugX — это открытая отладочная платформа, которая организует отладку LLM-агентов в замкнутый цикл «обнаружение — атрибуция — восстановление — повторный запуск». Ее ядро Deep...","url":"https://www.aioga.com/ru/news/cmrvnxfx7047nbihbzr6crygg/","contentTranslated":true,"sourceHash":"3316b9dd99dbe3c6","translatedAt":"2026-07-22T17:09:07.227Z"},"ar":{"title":"AgentDebugX: إطار عمل مفتوح المصدر لتصحيح الأخطاء لوكلاء الذكاء الاصطناعي LLM","summary":"AgentDebugX هو إطار عمل تصحيحي مفتوح المصدر، ينظم تصحيح وكيل LLM على شكل حلقة مغلقة \"الكشف - الإسناد - الاستعادة - إعادة التشغيل\". جوهره DeepDebug يحقق على معيار Who&When دقة في إسناد الوكلاء والخطوات لـ qwen3.5-9b، وعلى GAIA يمكنه إصلاح المهام الفاشلة بإعادة تشغيل واحدة فقط. توفر هذه الأداة مكتبة Python، وCLI، ولوحة تحكم ويب، ومهارات وكلاء قابلة للتثبيت.","category":"产品更新","source":"HuggingFace Daily Papers（社区热门论文）","aggregationSource":"HuggingFace Daily Papers（社区热门论文）","pageTitle":"AgentDebugX: إطار عمل مفتوح المصدر لتصحيح الأخطاء لوكلاء الذكاء الاصطناعي LLM - Aioga أخبار الذكاء الاصطناعي","description":"AgentDebugX هو إطار عمل تصحيحي مفتوح المصدر، ينظم تصحيح وكيل LLM على شكل حلقة مغلقة \"الكشف - الإسناد - الاستعادة - إعادة التشغيل\". جوهره DeepDebug يحقق على معيار Who&When دقة في إس...","url":"https://www.aioga.com/ar/news/cmrvnxfx7047nbihbzr6crygg/","contentTranslated":true,"sourceHash":"3316b9dd99dbe3c6","translatedAt":"2026-07-22T17:09:53.066Z"},"hi":{"title":"AgentDebugX：LLM एजेंटों के लिए ओपन-सोर्स फॉल्ट डिबगिंग फ्रेमवर्क","summary":"AgentDebugX एक ओपन-सोर्स डिबगिंग फ्रेमवर्क है, जो LLM एजेंट डिबगिंग को \"डिटेक्ट-एट्रिब्यूट-रिकवर-रीरन\" बंद लूप के रूप में व्यवस्थित करता है। इसका कोर DeepDebug Who&When बेंचमार्क पर qwen3.5-9b के लिए विशिष्ट एजेंट और स्टेप एट्रिब्यूशन सटीकता तक पहुँचता है, और GAIA पर एकल रीरन के द्वारा असफल कार्यों को ठीक कर सकता है। यह उपकरण Python लाइब्रेरी, CLI, वेब कंसोल और इंस्टॉल करने योग्य एजेंट क्षमताएँ प्रदान करता है।","category":"产品更新","source":"HuggingFace Daily Papers（社区热门论文）","aggregationSource":"HuggingFace Daily Papers（社区热门论文）","pageTitle":"AgentDebugX：LLM एजेंटों के लिए ओपन-सोर्स फॉल्ट डिबगिंग फ्रेमवर्क - Aioga AI समाचार","description":"AgentDebugX एक ओपन-सोर्स डिबगिंग फ्रेमवर्क है, जो LLM एजेंट डिबगिंग को \"डिटेक्ट-एट्रिब्यूट-रिकवर-रीरन\" बंद लूप के रूप में व्यवस्थित करता है। इसका कोर DeepDebug Who&When बेंचमार्क प...","url":"https://www.aioga.com/hi/news/cmrvnxfx7047nbihbzr6crygg/","contentTranslated":true,"sourceHash":"3316b9dd99dbe3c6","translatedAt":"2026-07-22T17:09:56.953Z"},"it":{"title":"AgentDebugX: un framework open source per il debug intelligente degli agenti LLM","summary":"AgentDebugX è un framework di debug open source che organizza il debug degli agenti LLM come un ciclo chiuso di \"rilevamento-attribuzione-recupero-riedizione\". Il suo nucleo DeepDebug raggiunge una precisione di attribuzione di agenti e passaggi su qwen3.5-9b secondo il benchmark Who&When, e su GAIA può correggere compiti falliti con una singola riedizione. Lo strumento fornisce librerie Python, CLI, console web e abilità agenti installabili.","category":"产品更新","source":"HuggingFace Daily Papers（社区热门论文）","aggregationSource":"HuggingFace Daily Papers（社区热门论文）","pageTitle":"AgentDebugX: un framework open source per il debug intelligente degli agenti LLM - Aioga Notizie IA","description":"AgentDebugX è un framework di debug open source che organizza il debug degli agenti LLM come un ciclo chiuso di \"rilevamento-attribuzione-recupero-riedizione\". Il suo nucleo DeepDe...","url":"https://www.aioga.com/it/news/cmrvnxfx7047nbihbzr6crygg/","contentTranslated":true,"sourceHash":"3316b9dd99dbe3c6","translatedAt":"2026-07-22T17:10:48.785Z"},"nl":{"title":"AgentDebugX: Open-source foutopsporingsframework voor LLM-agenten","summary":"AgentDebugX is een open-source debugframework dat LLM-agentdebugging organiseert in een 'detecteer-toereken-herstel-herstart' gesloten lus. De kern, DeepDebug, bereikt een nauwkeurige agent- en stap-toerekeningsnauwkeurigheid voor qwen3.5-9b op de Who&When-benchmark, en kan op GAIA mislukte taken met één enkele herstart repareren. De tool biedt een Python-bibliotheek, CLI, webconsole en installeerbare agentvaardigheden.","category":"产品更新","source":"HuggingFace Daily Papers（社区热门论文）","aggregationSource":"HuggingFace Daily Papers（社区热门论文）","pageTitle":"AgentDebugX: Open-source foutopsporingsframework voor LLM-agenten - Aioga AI-nieuws","description":"AgentDebugX is een open-source debugframework dat LLM-agentdebugging organiseert in een 'detecteer-toereken-herstel-herstart' gesloten lus. De kern, DeepDebug, bereikt een nauwkeur...","url":"https://www.aioga.com/nl/news/cmrvnxfx7047nbihbzr6crygg/","contentTranslated":true,"sourceHash":"3316b9dd99dbe3c6","translatedAt":"2026-07-22T17:10:40.648Z"},"tr":{"title":"AgentDebugX: LLM ajanlarına yönelik açık kaynaklı hata ayıklama çerçevesi","summary":"AgentDebugX, LLM ajanlarının hata ayıklamasını 'tespit-etkiden dolayı-onar-yeniden çalıştır' döngüsü olarak organize eden açık kaynaklı bir hata ayıklama çerçevesidir. Merkezi DeepDebug, Who&When kıyaslamasında qwen3.5-9b için doğru ajan ve adım atıf doğruluğuna ulaşır ve GAIA üzerinde tek bir yeniden çalıştırmayla başarısız görevleri düzeltebilir. Bu araç, Python kütüphanesi, CLI, Web kontrol paneli ve kurulum yapılabilir ajan becerileri sunar.","category":"产品更新","source":"HuggingFace Daily Papers（社区热门论文）","aggregationSource":"HuggingFace Daily Papers（社区热门论文）","pageTitle":"AgentDebugX: LLM ajanlarına yönelik açık kaynaklı hata ayıklama çerçevesi - Aioga AI Haberleri","description":"AgentDebugX, LLM ajanlarının hata ayıklamasını 'tespit-etkiden dolayı-onar-yeniden çalıştır' döngüsü olarak organize eden açık kaynaklı bir hata ayıklama çerçevesidir. Merkezi Deep...","url":"https://www.aioga.com/tr/news/cmrvnxfx7047nbihbzr6crygg/","contentTranslated":true,"sourceHash":"3316b9dd99dbe3c6","translatedAt":"2026-07-22T17:11:29.695Z"},"vi":{"title":"AgentDebugX: Khung gỡ lỗi mã nguồn mở dành cho tác nhân thông minh LLM","summary":"AgentDebugX là một framework gỡ lỗi mã nguồn mở, tổ chức việc gỡ lỗi các agent LLM theo vòng lặp khép kín 'Phát hiện - Phân tích nguyên nhân - Khôi phục - Chạy lại'. DeepDebug lõi của nó đạt độ chính xác cao trong việc phân tích nguyên nhân của agent và các bước trên benchmark Who&When với qwen3.5-9b, trên GAIA có thể sửa các nhiệm vụ thất bại chỉ với một lần chạy lại. Công cụ này cung cấp thư viện Python, CLI, bảng điều khiển Web và các kỹ năng agent có thể cài đặt.","category":"产品更新","source":"HuggingFace Daily Papers（社区热门论文）","aggregationSource":"HuggingFace Daily Papers（社区热门论文）","pageTitle":"AgentDebugX: Khung gỡ lỗi mã nguồn mở dành cho tác nhân thông minh LLM - Tin tức AI Aioga","description":"AgentDebugX là một framework gỡ lỗi mã nguồn mở, tổ chức việc gỡ lỗi các agent LLM theo vòng lặp khép kín 'Phát hiện - Phân tích nguyên nhân - Khôi phục - Chạy lại'. DeepDebug lõi...","url":"https://www.aioga.com/vi/news/cmrvnxfx7047nbihbzr6crygg/","contentTranslated":true,"sourceHash":"3316b9dd99dbe3c6","translatedAt":"2026-07-22T17:11:34.833Z"},"id":{"title":"AgentDebugX: Kerangka Debuging Sumber Terbuka untuk Agen Cerdas LLM","summary":"AgentDebugX adalah kerangka kerja debugging sumber terbuka yang mengorganisir debugging agen LLM menjadi siklus tertutup 'deteksi-atribusi-pemulihan-ulang'. DeepDebug intinya mencapai akurasi atribusi agen dan langkah yang tepat pada benchmark Who&When untuk qwen3.5-9b, dan pada GAIA, kegagalan tugas dapat diperbaiki hanya dengan satu kali pengulangan. Alat ini menyediakan pustaka Python, CLI, konsol web, dan kemampuan agen yang dapat diinstal.","category":"产品更新","source":"HuggingFace Daily Papers（社区热门论文）","aggregationSource":"HuggingFace Daily Papers（社区热门论文）","pageTitle":"AgentDebugX: Kerangka Debuging Sumber Terbuka untuk Agen Cerdas LLM - Berita AI Aioga","description":"AgentDebugX adalah kerangka kerja debugging sumber terbuka yang mengorganisir debugging agen LLM menjadi siklus tertutup 'deteksi-atribusi-pemulihan-ulang'. DeepDebug intinya menca...","url":"https://www.aioga.com/id/news/cmrvnxfx7047nbihbzr6crygg/","contentTranslated":true,"sourceHash":"3316b9dd99dbe3c6","translatedAt":"2026-07-22T17:12:15.288Z"},"th":{"title":"AgentDebugX: เฟรมเวิร์กการดีบักข้อผิดพลาดแบบโอเพนซอร์สสำหรับเอเจนต์ LLM","summary":"AgentDebugX เป็นเฟรมเวิร์กดีบักแบบโอเพนซอร์ส จัดการการดีบักเอเจนต์ LLM เป็นวงจรปิดแบบ \"ตรวจจับ-หาสาเหตุ-กู้คืน-รันใหม่\" ส่วนสำคัญ DeepDebug บนเกณฑ์ Who&When ทำให้ qwen3.5-9b สามารถระบุเอเจนต์และขั้นตอนอย่างแม่นยำ ใน GAIA การรันใหม่เพียงครั้งเดียวก็สามารถแก้ไขงานที่ล้มเหลว เครื่องมือนี้มีไลบรารี Python, CLI, เว็บคอนโซล และทักษะเอเจนต์ที่ติดตั้งได้","category":"产品更新","source":"HuggingFace Daily Papers（社区热门论文）","aggregationSource":"HuggingFace Daily Papers（社区热门论文）","pageTitle":"AgentDebugX: เฟรมเวิร์กการดีบักข้อผิดพลาดแบบโอเพนซอร์สสำหรับเอเจนต์ LLM - ข่าว AI Aioga","description":"AgentDebugX เป็นเฟรมเวิร์กดีบักแบบโอเพนซอร์ส จัดการการดีบักเอเจนต์ LLM เป็นวงจรปิดแบบ \"ตรวจจับ-หาสาเหตุ-กู้คืน-รันใหม่\" ส่วนสำคัญ DeepDebug บนเกณฑ์ Who&When ทำให้ qwen3.5-9b สามารถ...","url":"https://www.aioga.com/th/news/cmrvnxfx7047nbihbzr6crygg/","contentTranslated":true,"sourceHash":"3316b9dd99dbe3c6","translatedAt":"2026-07-22T17:12:24.598Z"},"pl":{"title":"AgentDebugX: Otwartoźródłowy framework do debugowania błędów dla agentów LLM","summary":"AgentDebugX to otwarta platforma do debugowania, która organizuje debugowanie agentów LLM w zamkniętym cyklu „wykrywanie-przypisywanie-przywracanie-powtórne uruchomienie”. Jej rdzeń DeepDebug osiąga na benchmarku Who&When dokładność przypisywania agenta i kroków dla qwen3.5-9b, a na GAIA pojedyncze powtórne uruchomienie może naprawić nieudane zadania. Narzędzie oferuje bibliotekę Pythona, CLI, konsolę internetową oraz instalowalne umiejętności agentów.","category":"产品更新","source":"HuggingFace Daily Papers（社区热门论文）","aggregationSource":"HuggingFace Daily Papers（社区热门论文）","pageTitle":"AgentDebugX: Otwartoźródłowy framework do debugowania błędów dla agentów LLM - Aioga Wiadomości AI","description":"AgentDebugX to otwarta platforma do debugowania, która organizuje debugowanie agentów LLM w zamkniętym cyklu „wykrywanie-przypisywanie-przywracanie-powtórne uruchomienie”. Jej rdze...","url":"https://www.aioga.com/pl/news/cmrvnxfx7047nbihbzr6crygg/","contentTranslated":true,"sourceHash":"3316b9dd99dbe3c6","translatedAt":"2026-07-22T17:13:13.094Z"}}}}