Overview

Jido Runic bridges Runic’s DAG-based workflow engine with Jido’s signal-driven agent framework, enabling durable, dataflow-driven agent workflows. Instead of imperative step-by-step orchestration, developers compose directed acyclic graphs (DAGs) of Jido Actions that execute according to dependency order — with full provenance tracking, signal gating, and automatic parallel dispatch.

The package solves a critical gap in agent design: how to express complex, multi-step pipelines (research, ETL, content generation, etc.) as composable, inspectable graphs that integrate natively with Jido’s signal routing, lifecycle hooks, and telemetry.

Purpose

Jido Runic is the workflow composition and execution substrate for durable DAG-based orchestration.

Boundary Lines

• Owns DAG workflow composition, scheduling, and provenance-aware execution semantics.
• Serves as orchestration substrate consumed by higher-level workflow packages.
• Does not own CLI provider adapters, runtime preflight/bootstrap policy, or session provisioning.

Major Components

ActionNode

Core adapter wrapping any Jido Action module as a Runic workflow node. Preserves stable content-addressed identity, introspects NimbleOptions schemas for input/output specs, and delegates execution to Jido.Exec.run/4.

SignalFact

Bidirectional adapter between Jido Signals and Runic Facts, converting causality tracking into fact ancestry chains and vice versa for continuous provenance.

SignalMatch

Runic match node gating downstream DAG execution based on Jido signal type prefix patterns for conditional workflow branches.

Strategy

Full Jido.Agent.Strategy implementation powered by a Runic DAG. Converts incoming signals to facts, emits runnables as directives, and automatically advances the workflow until satisfied.

Introspection

Provenance queries and execution summaries for walking fact ancestry chains and generating workflow state statistics.