Configuracion de MCP servers: project vs user scope, env-var expansion para credenciales y rotacion ante secreto leakeado.

Errores MCP tipados (isError, errorCategory, isRetryable, retryAfterSeconds); la politica de retry vive en el cliente.

Procesamiento masivo con Message Batches API, custom_id unico por request y fragmentacion selectiva de fallos parciales.

Propagacion de errores multi-agente: distinguir access failure de valid empty, local recovery primero y coverage gap annotation.

Prompt chaining multi-pass: pase local tipado por unidad y pase de integracion que solo ve resumenes, no las unidades crudas.

Reglas condicionales por glob de ruta; universales siempre cargadas, heuristicas de lenguaje on-demand.

Scratchpad persistente en disco curado por el agente; sobrevive a compact y reinicios, leido una vez y referenciado despues.

Exploracion segura en Plan Mode read-only con plan.md firmado por humano antes de transicionar a escritura.

Normalizacion de outputs heterogeneos via hook PostToolUse y updatedMCPToolOutput antes de entrar al historial del modelo.

Prompt caching contractual para Claude API: automático vs explícito, TTL, hasta cuatro breakpoints, retroceso, invalidación y verificación mediante usage.

Guardarrailes deterministas en hook PreToolUse con permissionDecision deny desde politica recargable, no en el system prompt.

Provenance tipada: no hay claim sin source; conflictos marcados con conflict true y escalados, nunca promediados.

Cross-cutting reliability patterns for PubNub apps. Covers reconnect with exponential backoff + jitter, idempotent publish with client-generated message IDs, dedup-on-merge for live + history streams, queue-and-retry for offline writes, and schema versioning of message envelopes. Use during design reviews, when planning offline support, or during incident response when network or delivery reliability is the concern.

Logging, testing, cost hygiene, incident triage, and usage metrics for PubNub apps. Covers the correlation fields every send/receive must log, the test pyramid for real-time apps, payload + fan-out cost hygiene, the incident triage runbook, and PubNub usage metrics for billing reconciliation. Use during code reviews, when planning monitoring, when triaging incidents, or when investigating PubNub cost overruns.

Manages PubNub apps, keysets, and environment separation. Covers the publish/subscribe/secret key model, dev/staging/prod isolation, key rotation hygiene, demo-key boundaries, and custom origin configuration. Use when setting up a new PubNub project, separating environments, rotating keys, configuring demo or production keysets, or asking about apps, keysets, or custom origins.

Builds real-time analytics and automation with PubNub Illuminate. Covers Business Objects (schema), Metrics (aggregations), Decisions (threshold-triggered actions with the 4-step PUT workflow), Queries (ad-hoc vs saved pipelines), and Dashboards. Use when tracking KPIs, building threshold alerts, automating mute/publish/App-Context-update actions, detecting spam or anomalies, or visualizing live activity.

Retrieves historical PubNub messages via Message Persistence (Storage & Playback). Covers timetoken-based pagination, per-channel ordering guarantees, offline catch-up flows, retention configuration, and the "catch-up tool not a data lake" principle. Use when fetching past messages, paginating with timetokens, building offline-resume UI, retrieving messages with actions, or configuring retention.

Routes PubNub events to external systems with no code via Events & Actions (E&A). Covers event listeners (Messages, Users, Channels, Push, Memberships), action targets (Webhook, SQS, Kinesis, S3, Kafka, IFTTT, AMQP), filter types (basic vs JSONPath), retry policy, envelopes, and batching. Use when integrating PubNub with Lambda, Kafka, SQS, S3, EventBridge, an analytics pipeline, or any external system.

Routes PubNub questions to the correct documentation source, MCP tool, and specialist skill. Classifies intent (chat vs non-chat, conceptual vs implementation, runtime testing vs analytics) and points the agent to the right next step. Use when a user mentions PubNub for the first time, asks "where do I start", "which docs", "what should I use", or any time the appropriate next skill is unclear.

Manages PubNub App Context (Objects API) for users, channels, and memberships. Covers user/channel metadata, custom fields, membership management, querying with filters, and referential integrity. Use when storing user profiles, channel metadata (name, topic, mute flags), tracking who is in which channel, or working with the Objects/App Context API.