Agentic System Design

The chapters in this module, in reading order.

#	Chapter
00	Designing agents — First-principles overview
01	Architecture decision tree — Which shape does this agent need?
02	The heartbeat that must stop — ReAct loop and the give-up rule
03	Contracts between model and world — Schemas that type the action, descriptions that route the intent
04	When tools chain and race — Parallel independence and serial handoff
05	One protocol, many servers — MCP as the standard memo between agent and world
06	What the agent forgets — Memory tiers, retrieval as tool, and the cost of remembering
07	What one bad call destroys — Blast radius classes and the approval gates that match
08	Budget before the first prompt — Cost, latency, and the multi-tenancy tax
09	Crash recovery and failure shapes — Checkpointing, idempotency, and topology-specific cracks
10	Alarm panel at build time — Observability and the eval gates that decide ship-or-no-ship
11	Shadow, canary, ramp, kill — Deploying agents and pinning what runs
12	The architect's checklist — your 20-item punch list before ship
13	Honest admission — Architecture decisions we still cannot defend at design time