[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"doc-\u002Far\u002Fdocs\u002Fsupervision\u002Fbenchmarks":3,"docs-nav-ar":469},{"_path":4,"_dir":5,"_draft":6,"_partial":6,"_locale":7,"title":8,"description":9,"body":10,"_type":463,"_id":464,"_source":465,"_file":466,"_stem":467,"_extension":468},"\u002Far\u002Fdocs\u002Fsupervision\u002Fbenchmarks","supervision",false,"","المعايير (Benchmarks)","قِس جودة الـpipeline بسيناريوهات قابلة لإعادة الإنتاج. اكتشف التراجعات قبل الإنتاج.",{"type":11,"children":12,"toc":454},"root",[13,36,42,47,67,72,78,83,106,112,268,273,279,304,333,343,349,368,373,405,411],{"type":14,"tag":15,"props":16,"children":17},"element","p",{},[18,21,27,29,34],{"type":19,"value":20},"text","الـ",{"type":14,"tag":22,"props":23,"children":24},"strong",{},[25],{"type":19,"value":26},"معيار (benchmark)",{"type":19,"value":28}," مجموعة من سيناريوهات التقييم تُعاد تشغيلها تلقائيًا على pipeline. وهو أداة قياس جودة الوكيل ",{"type":14,"tag":22,"props":30,"children":31},{},[32],{"type":19,"value":33},"بموضوعية",{"type":19,"value":35}," والتقاط التراجعات قبل أن تصل إلى الإنتاج.",{"type":14,"tag":37,"props":38,"children":40},"h2",{"id":39},"لماذا",[41],{"type":19,"value":39},{"type":14,"tag":15,"props":43,"children":44},{},[45],{"type":19,"value":46},"الوكيل الذي «بدا أنه يعمل» قد يتراجع بصمت بعد:",{"type":14,"tag":48,"props":49,"children":50},"ul",{},[51,57,62],{"type":14,"tag":52,"props":53,"children":54},"li",{},[55],{"type":19,"value":56},"تغيير في المهمّة؛",{"type":14,"tag":52,"props":58,"children":59},{},[60],{"type":19,"value":61},"تغيير في نموذج LLM؛",{"type":14,"tag":52,"props":63,"children":64},{},[65],{"type":19,"value":66},"إضافة أداة جديدة.",{"type":14,"tag":15,"props":68,"children":69},{},[70],{"type":19,"value":71},"بلا معيار، تكتشف ذلك في الإنتاج. ومع معيار، تراه في التشغيل التالي — قبل مستخدميك.",{"type":14,"tag":37,"props":73,"children":75},{"id":74},"بنية-السيناريو",[76],{"type":19,"value":77},"بنية السيناريو",{"type":14,"tag":15,"props":79,"children":80},{},[81],{"type":19,"value":82},"يحتوي السيناريو على:",{"type":14,"tag":48,"props":84,"children":85},{},[86,96],{"type":14,"tag":52,"props":87,"children":88},{},[89,94],{"type":14,"tag":22,"props":90,"children":91},{},[92],{"type":19,"value":93},"مُحفِّز (stimulus)",{"type":19,"value":95}," — الإدخال الذي يستقبله الـpipeline (رسالة أو تسلسل دورات، بيانات تجريبية fixtures).",{"type":14,"tag":52,"props":97,"children":98},{},[99,104],{"type":14,"tag":22,"props":100,"children":101},{},[102],{"type":19,"value":103},"فحوص (Checks)",{"type":19,"value":105}," — تأكيدات صريحة تُقيَّم على كل دورة: المخرَج المُنتَج، شكل الردّ، سلسلة استدعاءات الأدوات، تكلفة الإجراءات.",{"type":14,"tag":37,"props":107,"children":109},{"id":108},"كتالوج-الفحوص",[110],{"type":19,"value":111},"كتالوج الفحوص",{"type":14,"tag":113,"props":114,"children":115},"table",{},[116,135],{"type":14,"tag":117,"props":118,"children":119},"thead",{},[120],{"type":14,"tag":121,"props":122,"children":123},"tr",{},[124,130],{"type":14,"tag":125,"props":126,"children":127},"th",{},[128],{"type":19,"value":129},"الفحص",{"type":14,"tag":125,"props":131,"children":132},{},[133],{"type":19,"value":134},"ما يتحقّق منه",{"type":14,"tag":136,"props":137,"children":138},"tbody",{},[139,158,180,197,219,236],{"type":14,"tag":121,"props":140,"children":141},{},[142,153],{"type":14,"tag":143,"props":144,"children":145},"td",{},[146],{"type":14,"tag":147,"props":148,"children":150},"code",{"className":149},[],[151],{"type":19,"value":152},"pipeline_output_match",{"type":14,"tag":143,"props":154,"children":155},{},[156],{"type":19,"value":157},"مفتاح في تدفّق الإخراج يطابق القيمة المتوقَّعة (مسار + تأكيد)",{"type":14,"tag":121,"props":159,"children":160},{},[161,170],{"type":14,"tag":143,"props":162,"children":163},{},[164],{"type":14,"tag":147,"props":165,"children":167},{"className":166},[],[168],{"type":19,"value":169},"response_shape",{"type":14,"tag":143,"props":171,"children":172},{},[173,178],{"type":14,"tag":22,"props":174,"children":175},{},[176],{"type":19,"value":177},"شكل",{"type":19,"value":179}," ردّ الوكيل (الطول، البنية)، مُقاسًا دون حَكَم",{"type":14,"tag":121,"props":181,"children":182},{},[183,192],{"type":14,"tag":143,"props":184,"children":185},{},[186],{"type":14,"tag":147,"props":187,"children":189},{"className":188},[],[190],{"type":19,"value":191},"llm_rubric",{"type":14,"tag":143,"props":193,"children":194},{},[195],{"type":19,"value":196},"مضمون الردّ، يقيّمه حَكَم LLM وفق قائمة معايير قابلة للتحقّق",{"type":14,"tag":121,"props":198,"children":199},{},[200,209],{"type":14,"tag":143,"props":201,"children":202},{},[203],{"type":14,"tag":147,"props":204,"children":206},{"className":205},[],[207],{"type":19,"value":208},"tool_chain_analysis",{"type":14,"tag":143,"props":210,"children":211},{},[212,217],{"type":14,"tag":22,"props":213,"children":214},{},[215],{"type":19,"value":216},"تسلسل",{"type":19,"value":218}," استدعاءات الأدوات يتبع الاستراتيجية المتوقَّعة",{"type":14,"tag":121,"props":220,"children":221},{},[222,231],{"type":14,"tag":143,"props":223,"children":224},{},[225],{"type":14,"tag":147,"props":226,"children":228},{"className":227},[],[229],{"type":19,"value":230},"tool_cost",{"type":14,"tag":143,"props":232,"children":233},{},[234],{"type":19,"value":235},"ميزانية الإجراءات: لكل استدعاء تكلفة، وبلوغ الهدف يمنح مكافأة",{"type":14,"tag":121,"props":237,"children":238},{},[239,263],{"type":14,"tag":143,"props":240,"children":241},{},[242,248,250,256,257],{"type":14,"tag":147,"props":243,"children":245},{"className":244},[],[246],{"type":19,"value":247},"static",{"type":19,"value":249}," \u002F ",{"type":14,"tag":147,"props":251,"children":253},{"className":252},[],[254],{"type":19,"value":255},"tests",{"type":19,"value":249},{"type":14,"tag":147,"props":258,"children":260},{"className":259},[],[261],{"type":19,"value":262},"diff_quality",{"type":14,"tag":143,"props":264,"children":265},{},[266],{"type":19,"value":267},"فحوص موجَّهة للملفات والشيفرة (مخرَج pipeline تطوير)",{"type":14,"tag":15,"props":269,"children":270},{},[271],{"type":19,"value":272},"الكتالوج قابل للتوسّع على جانب المنصّة: كل فحص إضافة (plugin) مُسجَّلة على جانب الخادم.",{"type":14,"tag":37,"props":274,"children":276},{"id":275},"إنشاء-معيار",[277],{"type":19,"value":278},"إنشاء معيار",{"type":14,"tag":280,"props":281,"children":282},"ol",{},[283,293],{"type":14,"tag":52,"props":284,"children":285},{},[286,291],{"type":14,"tag":22,"props":287,"children":288},{},[289],{"type":19,"value":290},"التطبيقات ← التقييم ← السيناريوهات",{"type":19,"value":292},".",{"type":14,"tag":52,"props":294,"children":295},{},[296,298,303],{"type":19,"value":297},"أنشئ السيناريوهات (المُحفِّز، الفحوص، البيانات التجريبية)؛ كل سيناريو مخصّص لنموذج واحد (الـpipeline المستهدَف). وتُدار البيانات التجريبية في تبويب ",{"type":14,"tag":22,"props":299,"children":300},{},[301],{"type":19,"value":302},"الدعائم",{"type":19,"value":292},{"type":14,"tag":15,"props":305,"children":306},{},[307,309,316,318,324,325,331],{"type":19,"value":308},"يمكن أيضًا إنشاء سيناريو أو تعديله عبر مقترح من ",{"type":14,"tag":310,"props":311,"children":313},"a",{"href":312},"\u002Far\u002Fdocs\u002Fdeveloppeurs\u002Fmcp",[314],{"type":19,"value":315},"خادم MCP",{"type":19,"value":317}," (",{"type":14,"tag":147,"props":319,"children":321},{"className":320},[],[322],{"type":19,"value":323},"benchmark_create",{"type":19,"value":249},{"type":14,"tag":147,"props":326,"children":328},{"className":327},[],[329],{"type":19,"value":330},"benchmark_update",{"type":19,"value":332},")، يبقى معلّقًا إلى حين قبوله.",{"type":14,"tag":334,"props":335,"children":337},"alert",{"type":336},"info",[338],{"type":14,"tag":15,"props":339,"children":340},{},[341],{"type":19,"value":342},"المعيار الجيّد يتضمّن مزيجًا: الحالة الاسمية، الحالات الحدّية، الحالات العدائية (محاولات التلاعب، الالتباس، البيانات المفقودة).",{"type":14,"tag":37,"props":344,"children":346},{"id":345},"تشغيل-معيار",[347],{"type":19,"value":348},"تشغيل معيار",{"type":14,"tag":15,"props":350,"children":351},{},[352,354,359,361,366],{"type":19,"value":353},"من واجهة الإدارة، أطلِق ",{"type":14,"tag":22,"props":355,"children":356},{},[357],{"type":19,"value":358},"تشغيلًا",{"type":19,"value":360},": يوضَع في طابور متين ثم ينفّذه عامل (worker) مخصّص، خارج خادم الويب. وإذا توقّف هذا العامل في منتصف التشغيل، يُوسَم التشغيل بأنه ",{"type":14,"tag":22,"props":362,"children":363},{},[364],{"type":19,"value":365},"فاشل",{"type":19,"value":367}," عند إعادة التشغيل (ولا يُستأنَف تلقائيًا): أعِد إطلاقه. ويعرض درجٌ التفاصيل لكل سيناريو.",{"type":14,"tag":15,"props":369,"children":370},{},[371],{"type":19,"value":372},"يُظهِر التقرير:",{"type":14,"tag":48,"props":374,"children":375},{},[376,381,400],{"type":14,"tag":52,"props":377,"children":378},{},[379],{"type":19,"value":380},"معدّل النجاح لكل فحص",{"type":14,"tag":52,"props":382,"children":383},{},[384,386,391,393,398],{"type":19,"value":385},"تطوّر الدرجات بين عمليات التشغيل (تبويبا ",{"type":14,"tag":22,"props":387,"children":388},{},[389],{"type":19,"value":390},"السجلّ",{"type":19,"value":392}," و",{"type":14,"tag":22,"props":394,"children":395},{},[396],{"type":19,"value":397},"التطوّر",{"type":19,"value":399},")",{"type":14,"tag":52,"props":401,"children":402},{},[403],{"type":19,"value":404},"قائمة السيناريوهات الفاشلة، مع تفصيل كل فحص",{"type":14,"tag":37,"props":406,"children":408},{"id":407},"أفضل-الممارسات",[409],{"type":19,"value":410},"أفضل الممارسات",{"type":14,"tag":48,"props":412,"children":413},{},[414,424,434,444],{"type":14,"tag":52,"props":415,"children":416},{},[417,422],{"type":14,"tag":22,"props":418,"children":419},{},[420],{"type":19,"value":421},"ابدأ صغيرًا",{"type":19,"value":423}," — 5 سيناريوهات مُستهدَفة جيّدًا تساوي أكثر من 50 سيناريو تقريبيًا.",{"type":14,"tag":52,"props":425,"children":426},{},[427,432],{"type":14,"tag":22,"props":428,"children":429},{},[430],{"type":19,"value":431},"قِس قبل كل نشر",{"type":19,"value":433},"، وقارن بالتشغيل السابق قبل الدفع.",{"type":14,"tag":52,"props":435,"children":436},{},[437,442],{"type":14,"tag":22,"props":438,"children":439},{},[440],{"type":19,"value":441},"أخضِع مهامك لإدارة النسخ.",{"type":19,"value":443}," ليس الـpipelines فقط: فالمهام تتطوّر أيضًا.",{"type":14,"tag":52,"props":445,"children":446},{},[447,452],{"type":14,"tag":22,"props":448,"children":449},{},[450],{"type":19,"value":451},"حافظ على وضوح الفحوص.",{"type":19,"value":453}," الفحص الذي يتطلّب 20 سطرًا من الشرح لفهمه هو فحص يحتاج إلى تفكيك.",{"title":7,"searchDepth":455,"depth":455,"links":456},2,[457,458,459,460,461,462],{"id":39,"depth":455,"text":39},{"id":74,"depth":455,"text":77},{"id":108,"depth":455,"text":111},{"id":275,"depth":455,"text":278},{"id":345,"depth":455,"text":348},{"id":407,"depth":455,"text":410},"markdown","content:ar:docs:4.supervision:3.benchmarks.md","content","ar\u002Fdocs\u002F4.supervision\u002F3.benchmarks.md","ar\u002Fdocs\u002F4.supervision\u002F3.benchmarks","md",[470],{"title":471,"_path":472,"children":473},"Ar","\u002Far",[474],{"title":475,"_path":476,"children":477},"Docs","\u002Far\u002Fdocs",[478,491,519,547,558,574,586,647],{"title":479,"_path":480,"children":481},"البدء","\u002Far\u002Fdocs\u002Fdemarrer",[482,485,488],{"title":483,"_path":484},"مرحبًا بك","\u002Far\u002Fdocs\u002Fdemarrer\u002Fbienvenue",{"title":486,"_path":487},"إنشاء حساب","\u002Far\u002Fdocs\u002Fdemarrer\u002Fcreer-un-compte",{"title":489,"_path":490},"ربط مفتاح LLM","\u002Far\u002Fdocs\u002Fdemarrer\u002Fbrancher-llm",{"title":492,"_path":493,"children":494},"القنوات","\u002Far\u002Fdocs\u002Fcanaux",[495,498,501,504,507,510,513,516],{"title":496,"_path":497},"نظرة عامة على القنوات","\u002Far\u002Fdocs\u002Fcanaux\u002Fvue-densemble",{"title":499,"_path":500},"الهاتف في الوقت الفعلي","\u002Far\u002Fdocs\u002Fcanaux\u002Ftelephone",{"title":502,"_path":503},"البريد الإلكتروني","\u002Far\u002Fdocs\u002Fcanaux\u002Femail",{"title":505,"_path":506},"أداة دردشة الويب","\u002Far\u002Fdocs\u002Fcanaux\u002Fchat-widget",{"title":508,"_path":509},"HTTP Webhook","\u002Far\u002Fdocs\u002Fcanaux\u002Fwebhook",{"title":511,"_path":512},"المراسلة (WhatsApp، Messenger، Telegram، Mattermost)","\u002Far\u002Fdocs\u002Fcanaux\u002Fmessageries",{"title":514,"_path":515},"Meta Lead Ads","\u002Far\u002Fdocs\u002Fcanaux\u002Ffacebook-leads",{"title":517,"_path":518},"الجدولة (cron)","\u002Far\u002Fdocs\u002Fcanaux\u002Fschedule",{"title":520,"_path":521,"children":522},"Pipelines","\u002Far\u002Fdocs\u002Fpipelines",[523,526,529,532,535,538,541,544],{"title":524,"_path":525},"المحرّر المرئي","\u002Far\u002Fdocs\u002Fpipelines\u002Fediteur-visuel",{"title":527,"_path":528},"وكلاء الذكاء الاصطناعي","\u002Far\u002Fdocs\u002Fpipelines\u002Fagents",{"title":530,"_path":531},"المشغّلات الحتمية (Operators)","\u002Far\u002Fdocs\u002Fpipelines\u002Foperateurs",{"title":533,"_path":534},"التحكّم في التدفّق","\u002Far\u002Fdocs\u002Fpipelines\u002Fcontrole-flux",{"title":536,"_path":537},"خطوات الكود (Code Steps)","\u002Far\u002Fdocs\u002Fpipelines\u002Fcode-step",{"title":539,"_path":540},"الـPipelines الفرعية","\u002Far\u002Fdocs\u002Fpipelines\u002Fsous-pipelines",{"title":542,"_path":543},"إجراءات بلا كود","\u002Far\u002Fdocs\u002Fpipelines\u002Factions-no-code",{"title":545,"_path":546},"قاعدة المعرفة","\u002Far\u002Fdocs\u002Fpipelines\u002Fbase-de-connaissances",{"title":548,"_path":549,"children":550},"الإشراف","\u002Far\u002Fdocs\u002Fsupervision",[551,554,557],{"title":552,"_path":553},"سجلّ عمليات التنفيذ","\u002Far\u002Fdocs\u002Fsupervision\u002Fexecutions",{"title":555,"_path":556},"الأرصدة والقياس","\u002Far\u002Fdocs\u002Fsupervision\u002Fcredits-metering",{"title":8,"_path":4},{"title":559,"_path":560,"children":561},"الأمن","\u002Far\u002Fdocs\u002Fsecurite",[562,565,568,571],{"title":563,"_path":564},"مبادئ الأمن","\u002Far\u002Fdocs\u002Fsecurite\u002Fprincipes",{"title":566,"_path":567},"تعدّد المستأجرين الصارم","\u002Far\u002Fdocs\u002Fsecurite\u002Fmulti-tenant",{"title":569,"_path":570},"تدقيق GDPR وسجلّ الوصول","\u002Far\u002Fdocs\u002Fsecurite\u002Faudit-rgpd",{"title":572,"_path":573},"النشر السيادي","\u002Far\u002Fdocs\u002Fsecurite\u002Fdeploiement-souverain",{"title":575,"_path":576,"children":577},"المطوّرون","\u002Far\u002Fdocs\u002Fdeveloppeurs",[578,581,584],{"title":579,"_path":580},"واجهة برمجة تطبيقات الإدارة (Admin API)","\u002Far\u002Fdocs\u002Fdeveloppeurs\u002Fapi-admin",{"title":582,"_path":583},"webhooks الصادرة","\u002Far\u002Fdocs\u002Fdeveloppeurs\u002Fwebhooks-sortants",{"title":585,"_path":312},"قيادة betool من Claude (MCP)",{"title":587,"_path":588,"children":589},"التكاملات","\u002Far\u002Fdocs\u002Fintegrations",[590,593,596,599,602,605,608,611,614,617,620,623,626,629,632,635,638,641,644],{"title":591,"_path":592},"نظرة عامة على التكاملات","\u002Far\u002Fdocs\u002Fintegrations\u002Fvue-densemble",{"title":594,"_path":595},"Stripe","\u002Far\u002Fdocs\u002Fintegrations\u002Fstripe",{"title":597,"_path":598},"GitHub","\u002Far\u002Fdocs\u002Fintegrations\u002Fgithub",{"title":600,"_path":601},"Google","\u002Far\u002Fdocs\u002Fintegrations\u002Fgoogle",{"title":603,"_path":604},"LinkedIn Ads","\u002Far\u002Fdocs\u002Fintegrations\u002Flinkedin-ads",{"title":606,"_path":607},"Kavkom","\u002Far\u002Fdocs\u002Fintegrations\u002Fkavkom",{"title":609,"_path":610},"Jira","\u002Far\u002Fdocs\u002Fintegrations\u002Fjira",{"title":612,"_path":613},"Slack","\u002Far\u002Fdocs\u002Fintegrations\u002Fslack",{"title":615,"_path":616},"الرسائل القصيرة","\u002Far\u002Fdocs\u002Fintegrations\u002Fsms",{"title":618,"_path":619},"Enable Banking (الخدمات المصرفية المفتوحة)","\u002Far\u002Fdocs\u002Fintegrations\u002Fenable-banking",{"title":621,"_path":622},"CRM والتسويق والإنتاجية","\u002Far\u002Fdocs\u002Fintegrations\u002Fcrm-marketing",{"title":624,"_path":625},"HeyGen (الفيديو)","\u002Far\u002Fdocs\u002Fintegrations\u002Fheygen",{"title":627,"_path":628},"La Poste","\u002Far\u002Fdocs\u002Fintegrations\u002Fla-poste",{"title":630,"_path":631},"CRM متخصّص (beTool Application، Qhare)","\u002Far\u002Fdocs\u002Fintegrations\u002Fcrm-metier",{"title":633,"_path":634},"Mitti (الفحص الميداني)","\u002Far\u002Fdocs\u002Fintegrations\u002Fmitti",{"title":636,"_path":637},"Meta (Messenger، الصفحات، Meta Ads)","\u002Far\u002Fdocs\u002Fintegrations\u002Fmeta",{"title":639,"_path":640},"Onoff Business (Beta)","\u002Far\u002Fdocs\u002Fintegrations\u002Fonoff",{"title":642,"_path":643},"Smart Life \u002F Tuya (الأجهزة المتّصلة)","\u002Far\u002Fdocs\u002Fintegrations\u002Fsmart-life",{"title":645,"_path":646},"منتجات beTool (Desktop، Kiosk، Terminal)","\u002Far\u002Fdocs\u002Fintegrations\u002Fproduits-betool",{"title":648,"_path":649,"children":650},"حالات الاستخدام","\u002Far\u002Fdocs\u002Fcas-usage",[651,654,657,660,663,666],{"title":652,"_path":653},"agents أعمال نموذجية","\u002Far\u002Fdocs\u002Fcas-usage\u002Fvue-densemble",{"title":655,"_path":656},"CFO IA","\u002Far\u002Fdocs\u002Fcas-usage\u002Fcfo-ia",{"title":658,"_path":659},"COO IA","\u002Far\u002Fdocs\u002Fcas-usage\u002Fcoo-ia",{"title":661,"_path":662},"CTO IA","\u002Far\u002Fdocs\u002Fcas-usage\u002Fcto-ia",{"title":664,"_path":665},"Product Manager IA","\u002Far\u002Fdocs\u002Fcas-usage\u002Fproduct-manager-ia",{"title":667,"_path":668},"المساعد الشخصي","\u002Far\u002Fdocs\u002Fcas-usage\u002Fassistant-personnel"]