AGNTMOD.AGNT-07 · v1.0

Agents that take
action, not chatbots
that answer.

7 micro-lessons · ~66 min · Real Docker images

THE REACT LOOP · LIVE
SEQ 04/04 · 110 BPM
RUNNING
01
PERCEIVE
Read input + state
02
THINK
Plan next action
03
ACT
Call a tool
04
OBSERVE
Read tool result
LOOP · feed observation back as next perception
AGNTAI ENGINEERINGHOT

Agentic AI / AI agents

Build agents that take action, not chatbots that answer.

WHY THIS MATTERS · MCKINSEY · STATE OF AI 2026
23% of organizations are scaling agentic AI. Another 39% are experimenting with it.
WHAT YOU'LL LEARN
01Agent vs chatbot
02Tool calling basics
03ReAct reasoning loop
04Memory & state
05Multi-step planning
06Production guardrails
07Cost & latency budget
YOU'LL BE ABLE TO
Design tool-using agents end-to-end
Ship a ReAct loop with guardrails
Cost-bound an agent in production
SKILLS YOU'LL GAIN

Real skills, real career delta.

Skills you'll gain

08
  • Build the smallest agent that existsWorking

    Loop + tools = agent. Author the 12-line ReAct loop from memory and explain why it isn't a chatbot.

  • Author production-grade toolsProduction

    Tight JSON Schema + idempotent side effects + descriptions the model can route on. Three tools beat fifteen.

  • Trace and debug the ReAct loopProduction

    THINK before ACT, parse the model's reasoning, log ACT pre-dispatch — turn invisible failures into readable trace lines.

  • Wire memory & state correctlyWorking

    Facts → SQL, similarity → vector, conversation → msgs[]. Sliding-window N + similarity floor 0.7.

  • Plan-then-execute past 5 stepsProduction

    Persist the plan to disk; reactive ReAct fails past 5 hops. Plan-then-execute beats reactive at scale.

  • Bound cost in productionProduction

    Per-session token budgets, cheaper model for routing, cached deterministic tool results, Prom metric in dollars.

  • Layer agent safety end-to-endAdvanced

    Input filter → sandbox → output judge → audit log. Output judge runs on a SEPARATE model role. Adversarial inputs in CI.

  • Ship a guarded agent to productionProduction

    Tool schemas + step cap + audit log + budget cutoff + Prom-exposed metrics + a different-model output guard.

RUNNABLE ON YOUR MACHINE
$ docker pull snap/agentic-ai:lesson-01
$ docker run --rm -it snap/agentic-ai:lesson-01
snap/agentic-ai:lesson-01
QUICK PREVIEW · 7 MIN
VERIFIED ENGINEER REVIEWS
Finally an agents track that ships working code, not slideware.
@sre_mayaVERIFY ON GITHUB
The ReAct loop lesson clicked in 4 minutes.
@kofi.infraVERIFY ON TWITTER
LESSONS7
HOURS~1.1
LEARNERS4,128
THIS WEEK+24%