{"apiVersion":"v1","directorySlug":"devis-assurance-obseque-mcp-cf0e3db500","title":"Devis Assurance Obseque","description":null,"pageUrl":"https://www.mcpbundles.com/skills/devis-assurance-obseque-mcp-cf0e3db500","connectPath":"importable","connectAction":"import_server","catalogListing":false,"bundleSlug":null,"listingKind":"remote_mcp","kind":"provider","transportUrl":"https://devis-assurance-obseque.com/mcp","hostedMcpUrl":null,"toolCount":15,"toolInventoryStatus":"observed","refreshedAt":"2026-11-03T13:26:00.516172+00:00","registryName":"com.devis-assurance-obseque/devis-assurance-obseque-mcp-cf0e3db500","registryVersion":"2026.11.3","registryEligible":true,"tools":[{"name":"list_tests","description":"Liste tous les tests A/B avec leur configuration. Retourne pour chaque test : test_id, name, status, url_from (pattern URL), optimize (on/off), discovery_rate (taux exploration epsilon-greedy), nombre de variations et liste des noms de variations. Utile pour decouvrir les tests disponibles avant d'utiliser les autres outils.","inputSchema":{"type":"object","properties":{"limit":{"type":"integer"},"status":{"enum":["on","off","all"],"type":"string"}}},"logicalName":"list_tests","displayName":"list_tests","title":"list_tests","annotations":{"untrustedContentHint":true,"readOnlyHint":true}},{"name":"get_test_details","description":"Recupere le detail complet d'un test A/B par son test_id. Retourne : configuration du test, liste des variations avec leurs scores PageSpeed (desktop/mobile), resume des keyword groups (clustering IA), et configuration des goals. Les variations sont liees par STRING (variation_name), pas par variation_id.","inputSchema":{"type":"object","required":["test_id"],"properties":{"test_id":{"type":"integer"}}},"logicalName":"get_test_details","displayName":"get_test_details","title":"get_test_details","annotations":{"untrustedContentHint":true,"readOnlyHint":true}},{"name":"get_statistics","description":"Requete statistique agregee sur les visites (visit_log), groupee par une dimension au choix. C'est l'outil principal pour analyser les performances. Dimensions disponibles : variation, keyword, keyword_group, utm_source, utm_campaign, device_type, os, browser, uri, uri_path, abtest, optimization_method. Metriques retournees : visits, views, goals, goal_rate (%), phone_validation, phone_validation_rate (%), bounce_rate (%), avg_duration (s), avg_server_response_time (ms). TIMESTAMPS : start_date/","inputSchema":{"type":"object","required":["start_date","end_date"],"properties":{"filters":{"type":"object","properties":{"os":{"type":["string","array"]},"device":{"type":["string","array"]},"browser":{"type":["string","array"]},"keywords":{"type":["string","array"]},"variation":{"type":["string","array"]},"utm_source":{"type":["string","array"]},"uri_contains":{"type":"string"},"utm_campaign":{"type":["string","array"]},"keyword_group":{"type":["string","array"]}}},"test_id":{"type":["integer","array"]},"end_date":{"type":"string"},"dimension":{"enum":["variation","keyword","keyword_group","utm_source","utm_campaign","device_type","os","browser","uri","uri_path","abtest","optimization_method"],"type":"string"},"start_date":{"type":"string"},"explode_optimization":{"type":"boolean"}}},"logicalName":"get_statistics","displayName":"get_statistics","title":"get_statistics","annotations":{"untrustedContentHint":true,"readOnlyHint":true}},{"name":"get_daily_trends","description":"Retourne les series temporelles jour par jour pour analyse de tendances. Memes parametres que get_statistics mais les resultats sont ventiles par date. Utile pour detecter des tendances, saisonnalite, regressions de performance. TIMESTAMPS : start_date/end_date acceptent ISO 8601 (YYYY-MM-DD) ou UNIX epoch.","inputSchema":{"type":"object","required":["start_date","end_date"],"properties":{"filters":{"type":"object","properties":{"os":{"type":["string","array"]},"device":{"type":["string","array"]},"browser":{"type":["string","array"]},"keywords":{"type":["string","array"]},"variation":{"type":["string","array"]},"utm_source":{"type":["string","array"]},"utm_campaign":{"type":["string","array"]}}},"test_id":{"type":["integer","array"]},"end_date":{"type":"string"},"dimension":{"enum":["variation","keyword","keyword_group","utm_source","utm_campaign","device_type","os","browser","uri","uri_path","abtest","optimization_method"],"type":"string"},"start_date":{"type":"string"}}},"logicalName":"get_daily_trends","displayName":"get_daily_trends","title":"get_daily_trends","annotations":{"untrustedContentHint":true,"readOnlyHint":true}},{"name":"get_variation_comparison","description":"Comparaison directe de 2+ variations d'un test A/B. Calcule pour chaque variation : visits, goals, goal_rate, phone_validation_rate, bounce_rate, avg_duration. Determine le winner avec le lift % par rapport au baseline (premiere variation = original). Calcule la confiance statistique (Z-test sur proportions). Genere une recommandation textuelle. TIMESTAMPS : start_date/end_date acceptent ISO 8601 (YYYY-MM-DD) ou UNIX epoch.","inputSchema":{"type":"object","required":["test_id"],"properties":{"filters":{"type":"object","properties":{"device":{"type":["string","array"]},"keywords":{"type":["string","array"]},"utm_source":{"type":["string","array"]},"utm_campaign":{"type":["string","array"]}}},"test_id":{"type":"integer"},"end_date":{"type":"string"},"start_date":{"type":"string"},"variations":{"type":"array","items":{"type":"string"}}}},"logicalName":"get_variation_comparison","displayName":"get_variation_comparison","title":"get_variation_comparison","annotations":{"untrustedContentHint":true,"readOnlyHint":true}},{"name":"get_funnel_analysis","description":"Analyse de funnel profonde pour une variation donnee. Retourne : engagement (goal_rate, bounce_rate, duree moyenne), segmentation des abandons (bounce / soft_bounce / early / qualified / late / converted), taux de completion par etape de formulaire (form_steps), points de friction (champs ou les visiteurs abandonnent), comparaison convertis vs abandonnes (duree, profondeur, champs remplis). Utilise les donnees form_steps JSON de visit_log pour l'analyse formulaire. TIMESTAMPS : start_date/end_da","inputSchema":{"type":"object","required":["variation"],"properties":{"limit":{"type":"integer"},"filters":{"type":"object","properties":{"os":{"type":["string","array"]},"device":{"type":["string","array"]},"browser":{"type":["string","array"]},"keywords":{"type":["string","array"]},"utm_source":{"type":["string","array"]},"utm_campaign":{"type":["string","array"]}}},"end_date":{"type":"string"},"variation":{"type":"string"},"start_date":{"type":"string"}}},"logicalName":"get_funnel_analysis","displayName":"get_funnel_analysis","title":"get_funnel_analysis","annotations":{"untrustedContentHint":true,"readOnlyHint":true}},{"name":"get_web_vitals","description":"Analyse des Core Web Vitals : donnees RUM reelles (visit_log) + scores Lighthouse lab (variation). RUM : LCP, FCP, CLS, TTFB moyennes et P75 par variation, avec rating (good/needs-improvement/poor). Lighthouse : scores PageSpeed desktop/mobile depuis la table variation. Compte les pages lentes (score composite < 90). Permet de correler performance technique et taux de conversion. TIMESTAMPS : start_date/end_date acceptent ISO 8601 ou UNIX epoch.","inputSchema":{"type":"object","required":["start_date","end_date"],"properties":{"device":{"type":"string"},"test_id":{"type":"integer"},"end_date":{"type":"string"},"variation":{"type":"string"},"start_date":{"type":"string"}}},"logicalName":"get_web_vitals","displayName":"get_web_vitals","title":"get_web_vitals","annotations":{"untrustedContentHint":true,"readOnlyHint":true}},{"name":"get_best_performance","description":"Simule la recherche hierarchique de la meilleure variation telle que le moteur la fait en production. Utilise la table performance (agregation pre-calculee) avec 6 niveaux de granularite : Niveau 1: test_id + variation + device_type + time_slot + utm_source + utm_campaign + keywords. Niveau 2: test_id + variation + device_type + time_slot + utm_source + utm_campaign. Niveau 3: test_id + variation + device_type + time_slot + utm_source. Niveau 4: test_id + variation + device_type + time_slot. Niv","inputSchema":{"type":"object","required":["test_id"],"properties":{"test_id":{"type":"integer"},"keywords":{"type":"string"},"time_slot":{"type":"string"},"timestamp":{"type":"integer"},"utm_source":{"type":"string"},"device_type":{"type":"string"},"utm_campaign":{"type":"string"}}},"logicalName":"get_best_performance","displayName":"get_best_performance","title":"get_best_performance","annotations":{"untrustedContentHint":true,"readOnlyHint":true}},{"name":"get_keyword_performance","description":"Analyse de performance par keyword brut ou par keyword_group (clustering IA Claude). Retourne pour chaque keyword/groupe : visits, goals, goal_rate, phone_validation_rate, bounce_rate, avg_duration. En mode keyword_group, indique aussi le nombre de keywords dans chaque groupe et les keywords non-classes. Jointure logique : keyword_group.keyword_raw = visit_log.keywords. TIMESTAMPS : start_date/end_date acceptent ISO 8601 ou UNIX epoch.","inputSchema":{"type":"object","required":["test_id","start_date","end_date"],"properties":{"test_id":{"type":"integer"},"end_date":{"type":"string"},"group_by":{"enum":["keyword","keyword_group"],"type":"string"},"min_visits":{"type":"integer"},"start_date":{"type":"string"}}},"logicalName":"get_keyword_performance","displayName":"get_keyword_performance","title":"get_keyword_performance","annotations":{"untrustedContentHint":true,"readOnlyHint":true}},{"name":"get_traffic_sources","description":"Analyse de performance par source de trafic (utm_source ou utm_campaign). Option cross_with_variation pour croiser source x variation et identifier quelles sources marchent avec quelles landings. Retourne visits, goals, goal_rate, phone_validation_rate, bounce_rate, avg_duration par source. TIMESTAMPS : start_date/end_date acceptent ISO 8601 ou UNIX epoch.","inputSchema":{"type":"object","required":["start_date","end_date"],"properties":{"test_id":{"type":"integer"},"end_date":{"type":"string"},"group_by":{"enum":["utm_source","utm_campaign"],"type":"string"},"start_date":{"type":"string"},"cross_with_variation":{"type":"boolean"}}},"logicalName":"get_traffic_sources","displayName":"get_traffic_sources","title":"get_traffic_sources","annotations":{"untrustedContentHint":true,"readOnlyHint":true}},{"name":"query_visits","description":"Requete sur les visites individuelles (visit_log) avec tous les champs disponibles. Retourne les visites brutes avec : visitId, timestamp, uri, variation, test_id, device_type, time_slot (matin/apres-midi/soir/nuit, derive du timestamp via Visit::computeTimeSlot), keywords, utm_source, utm_campaign, goal_nb, phone_validation, visit_duration, form_steps (JSON), optimization_method (explore/exploit/control), decision_id (référence vers le journal du moteur A/B → à passer à explain_decision pour le","inputSchema":{"type":"object","properties":{"limit":{"type":"integer"},"device":{"type":"string"},"test_id":{"type":"integer"},"end_date":{"type":"string"},"has_goal":{"type":"boolean"},"keywords":{"type":"string"},"order_by":{"enum":["visitId","timestamp","visit_duration","goal_nb"],"type":"string"},"order_dir":{"enum":["asc","desc"],"type":"string"},"time_slot":{"enum":["matin","apres-midi","soir","nuit"],"type":"string"},"variation":{"type":"string"},"start_date":{"type":"string"},"utm_source":{"type":"string"},"utm_campaign":{"type":"string"},"slow_pages_only":{"type":"boolean"}}},"logicalName":"query_visits","displayName":"query_visits","title":"query_visits","annotations":{"untrustedContentHint":true,"readOnlyHint":true}},{"name":"query_database","description":"Execute une requete SQL SELECT en lecture seule sur la base de donnees A/B Testing. Filet de securite pour les questions imprevues que les autres outils ne couvrent pas. TABLES DISPONIBLES (7) : test, visit_log, performance, variation, goal, keyword_group, redirection. IMPORTANT : tous les timestamps sont en UNIX epoch (INT secondes depuis 1970-01-01), sauf goal.created_at (MYSQL TIMESTAMP). Utiliser FROM_UNIXTIME(timestamp) pour convertir en date lisible. Utiliser DATE(FROM_UNIXTIME(timestamp))","inputSchema":{"type":"object","required":["query"],"properties":{"limit":{"type":"integer"},"query":{"type":"string"}}},"logicalName":"query_database","displayName":"query_database","title":"query_database","annotations":{"untrustedContentHint":true,"readOnlyHint":true}},{"name":"explain_decision","description":"Explique la décision du moteur A/B test pour UNE visite donnée : la bonne variation a-t-elle été choisie, et si non pourquoi. Croise visit_log (servi + vraies dimensions), abtest.log (raisonnement exact figé au moment de la décision : dimensions réellement utilisées, cascade niveau par niveau, explore/exploit), et un replay live. Détecte : dimension connue mais perdue par le moteur, choix sur segment trop faible (< seuil de visites), repli aléatoire (fallback), exploitation sur peu de dimensions","inputSchema":{"type":"object","properties":{"replay":{"type":"boolean"},"visit_id":{"type":"integer"},"decision_id":{"type":"string"}}},"logicalName":"explain_decision","displayName":"explain_decision","title":"explain_decision","annotations":{"untrustedContentHint":true,"readOnlyHint":true}},{"name":"verify_aggregates","description":"Vérifie que la table performance est correctement calculée : recompute l'agrégat attendu depuis visit_log (fenêtre glissante) et le diffe contre la table. Détecte valeurs fausses (agrégat périmé), lignes manquantes, lignes périmées. Paramètre level (0=toutes dimensions … 5=totaux par variation, défaut 5, le plus fiable). Dimensions du moteur dans l'ordre : device_type, time_slot, utm_source, utm_campaign, keywords. Idéal après explain_decision quand on suspecte un mauvais calcul de performance.","inputSchema":{"type":"object","required":["test_id"],"properties":{"level":{"type":"integer"},"limit":{"type":"integer"},"test_id":{"type":"integer"},"tolerance":{"type":"integer"},"variation":{"type":"string"},"window_days":{"type":"integer"}}},"logicalName":"verify_aggregates","displayName":"verify_aggregates","title":"verify_aggregates","annotations":{"untrustedContentHint":true,"readOnlyHint":true}},{"name":"analyze_dimensions","description":"Recommande quelles dimensions d'optimisation activer pour un test A/B (lesquelles, dans quel ordre), OÙ arrêter la granularité, et CHIFFRE le gain : uplift de conversion de la config recommandée vs la config actuelle, rejoué hors-échantillon sur les N derniers jours (replay off-policy doublement robuste, IC95 par bootstrap). Verdict APPLIQUER / NE RIEN CHANGER selon que l'IC95 exclut 0. Fenêtres train/test auto (en jours, maximisées sous contrainte de rotation des variations). Champ « rapport »","inputSchema":{"type":"object","required":["test_id"],"properties":{"algo":{"enum":["lcb","greedy"],"type":"string"},"test_id":{"type":"integer"},"min_gain":{"type":"number"},"min_goals":{"type":"integer"},"test_window_days":{"type":"integer"},"train_window_days":{"type":"integer"}}},"logicalName":"analyze_dimensions","displayName":"analyze_dimensions","title":"analyze_dimensions","annotations":{"untrustedContentHint":true,"readOnlyHint":true}}],"toolSource":"directory_preview","agentAction":"both","installLinks":[{"id":"cursor","label":"Cursor","href":"cursor://anysphere.cursor-deeplink/mcp/install?name=Devis-Assurance-Obseque&config=eyJ1cmwiOiAiaHR0cHM6Ly9kZXZpcy1hc3N1cmFuY2Utb2JzZXF1ZS5jb20vbWNwIn0%3D"},{"id":"vscode","label":"VS Code","href":"vscode:mcp/install?%7B%22name%22%3A%22Devis-Assurance-Obseque%22%2C%22type%22%3A%22http%22%2C%22url%22%3A%22https%3A%2F%2Fdevis-assurance-obseque.com%2Fmcp%22%7D"},{"id":"vscode-insiders","label":"VS Code Insiders","href":"vscode-insiders:mcp/install?%7B%22name%22%3A%22Devis-Assurance-Obseque%22%2C%22type%22%3A%22http%22%2C%22url%22%3A%22https%3A%2F%2Fdevis-assurance-obseque.com%2Fmcp%22%7D"},{"id":"claude-web","label":"Claude.ai","href":"https://claude.ai/customize/connectors?modal=add-custom-connector&connectorName=Devis+Assurance+Obseque&connectorUrl=https%3A%2F%2Fdevis-assurance-obseque.com%2Fmcp"},{"id":"codex","label":"Codex","copyText":"codex mcp add \"Devis-Assurance-Obseque\" --url \"https://devis-assurance-obseque.com/mcp\""},{"id":"chatgpt","label":"ChatGPT","href":"https://www.mcpbundles.com/docs/integrations/chatgpt"}],"catalogUrl":"https://www.mcpbundles.com/api/directory/v1/servers/devis-assurance-obseque-mcp-cf0e3db500","callUrl":"https://www.mcpbundles.com/api/directory/v1/servers/devis-assurance-obseque-mcp-cf0e3db500/tools/{tool}","mcpServerUrl":"https://devis-assurance-obseque.com/mcp","liveCatalogUrl":"https://www.mcpbundles.com/api/directory/v1/servers/devis-assurance-obseque-mcp-cf0e3db500?live=1"}