Keyboard shortcuts

Press ← or → to navigate between chapters

Press S or / to search in the book

Press ? to show this help

Press Esc to hide this help

Agentic AI with Sruja

Learn to design agent-based AI systems with clear boundaries, interfaces, and governance using Sruja DSL.

Learning objectives

  • Model orchestration, tools, and memory in .sruja (architecture description—not runtime execution).
  • Distinguish Sruja as a grounded harness (lint, drift, evidence, MCP) from the editor/CI host that runs the LLM loop.
  • Operate continual learning in token space: agent memory, bounded agent plan / agent apply, and optional local inference via --enrich-cmd.

For how common multi-agent and MCP narratives map to Sruja’s scope (grounding vs. runtime orchestration), see Agentic orchestration patterns and Sruja. For harness + host-owned learning, see Grounded harness and continual learning. For the Context Graph model, custom schema syntax, and default validation rules, see Domain schema and context graphs.