Run durable AI-agent workflows on Dapr
Use Dapr sidecars and workflows to give AI agents durable execution, state, messaging, approvals, and recovery across languages and platforms.
npx skills add agentskillexchange/skills --skill run-durable-ai-agent-workflows-on-dapr
Use Dapr when an agentic application needs production runtime guarantees instead of a fragile in-process loop. The operator defines long-running agent or multi-agent workflows as normal application code, runs them with Dapr sidecars, and relies on Dapr for durable progress, state persistence, event-driven coordination, service identity, communication, observability, and recovery after crashes or restarts. A practical workflow is to install the Dapr CLI, initialize the runtime, wire the agent service to workflow, state, pub/sub, conversation, or service-invocation APIs, and deploy the same pattern locally or on Kubernetes. The boundary is runtime reliability for agent workflows and agent services; it is not a generic microservices platform card or a replacement for an agent framework’s reasoning layer.
What this skill actually does
Inputs and prerequisites: Dapr CLI, Docker or Kubernetes runtime, application language SDK or HTTP/gRPC client.
Setup notes: Install the Dapr CLI from the official Dapr docs, initialize a local or Kubernetes runtime, then run agent workflow services with Dapr sidecars and the workflow, state, pub/sub, conversation, or service-invocation APIs required by the application.
Source and verification boundary: use https://dapr.io as the canonical reference before running the workflow; keep commands, API calls, CLI usage, and generated outputs reviewable against that upstream source.
Framework fit: publish this as a Multi-Framework workflow only when the operator can invoke the documented toolchain directly, rather than treating the upstream project as a generic product listing.