5,814 agentes preconstruidos y 15,654 habilidades de biblioteca en 24 sectores y 12 roles por sector, instalables desde el catálogo, conectados a tus sistemas y ejecutados a través de la misma cadena de puertas de fallo cerrado que todo lo demás en Cortex.
21,511 plantillas de biblioteca en la matriz del catálogo
Suites creadas a mano para los trabajos que las empresas automatizan primero — cada colección se entrega gobernada, con políticas y auditoría incorporadas.
Busca entre 21,500+ habilidades y agentes gobernados, o filtra por sector y tipo: cada resultado se instala desde el catálogo y pasa por las mismas puertas.
402 resultados · Agentes · “HR”
Build comprehensive model monitoring dashboards tracking prediction quality, data drift, latency, throughput, and cost metrics with automated alerting and anomaly detection.
Continuously monitor model performance using discrimination, calibration, and stability metrics. Detects model drift, concept drift, and triggers revalidation when thresholds are breached.
End-to-end management of regulatory examinations from notice receipt through final response. Coordinates data requests, manages interview prep, tracks findings, and drafts formal responses.
Calibrate risk appetite statement metrics and tolerances. Translates board-level risk appetite into quantitative metrics, thresholds, and escalation triggers across risk categories.
Ensure data quality for regulatory reporting through automated profiling, anomaly detection, completeness checks, lineage tracing, and data quality scorecards.
Develop and deploy SIEM detection rules -- threat model input, rule logic creation, backtesting against historical data, false positive tuning, staged deployment, and detection validation.
Fine-tune semantic search relevance by optimizing embedding models, query expansion strategies, hybrid scoring weights, and reranking thresholds using relevance feedback.
AI-driven detection rule optimization -- analyze false positive rates, tune thresholds, test against historical data, validate detection coverage, and deploy optimized rules.
Design and execute regulatory and internal stress test scenarios. Defines multi-factor stress scenarios, runs them through actuarial and financial models, and quantifies capital impact.
Executes a proactive threat hunt using a hypothesis-driven approach. Maps to MITRE ATT&CK techniques, queries log sources, correlates IOCs, and documents findings.
Enriches IOCs (IPs, domains, hashes, URLs) with threat actor profiling, multi-source feed correlation, MITRE ATT&CK technique mapping, and confidence scoring.
Assess transaction monitoring coverage gaps by mapping existing rules to known typologies, regulatory expectations, and emerging threats. Identifies unmonitored risk areas.
Develop new transaction monitoring scenarios from typology research, regulatory guidance, and emerging threat intelligence. Includes rule specification and testing plan.
Analyze historical alert data to optimize transaction monitoring thresholds. Reduces false positives while maintaining regulatory detection coverage using statistical back-testing.
Analyze portfolio mix against appetite and steering targets, identify segments to grow/shrink.
Perform VaR backtesting using Kupiec POF, Christoffersen independence, and Basel traffic-light tests. Identifies model deficiencies and triggers recalibration alerts.
Working productivity agent. Turns supplied meeting context into a concise pre-read; it does not read calendars, CRM records, or message threads automatically.
Working supplied-context agent. Searches approved HR knowledge, summarizes supplied context, and drafts a policy-grounded response; it does not create or route cases.
Los agentes de la biblioteca no son scripts: son identidades gobernadas. Instalar uno lo registra con un propietario, un presupuesto, modelos y acciones permitidos y un lugar en el registro de auditoría.
Filtra por tu industria y tu rol; cada elemento indica los sistemas a los que se conecta y las habilidades que invoca.
Conectores, servidores MCP e importaciones OpenAPI — acotados por listas de permitidos y DLP en el gateway.
Identidad, presupuesto, guardrails, registro, Control Tower, ejecución, guarda de salida — y luego un recibo firmado en el Trust Ledger. En todos los planes.
El constructor visual de habilidades permite lanzar nuevas habilidades sin código — o dinos qué te falta y te señalaremos el patrón gobernado más cercano.