AI Agents cookbook

14 recipes · 92 steps · about 93 minutes of reading

Write a tool-using agent loop that knows when to stop
Every agent tutorial hides the loop inside somebody else's framework.
BeginnerRead 5 minDoing: 30 to 45 minutes with your API key and an editor open6 steps
One bad tool call crashes the whole agent run
Your agent calls a tool, the tool throws, and the whole script dies instead of the agent trying something else.
Beginner to intermediateRead 6 minDoing: 30 to 45 minutes6 steps
Log every tool call so you can find what broke a run
Your agent finished and handed back an answer that reads like it did exactly what you asked, but there is no record of which tool call actually produced it.
Beginner to intermediateRead 5 minDoing: 45 minutes to an hour5 steps
Split an overloaded prompt into a checked three-step chain
One prompt is asked to plan, draft, and check itself, and it quietly skips one of those jobs.
Beginner to intermediateRead 6 minDoing: an hour with your API key and an editor open7 steps
Give a runaway agent a spending cap
A stuck loop can burn through a month's API budget before anyone notices.
IntermediateRead 5 minDoing: 45 minutes to an hour6 steps
Shrink an agent's blast radius by scoping down its tools
You gave your agent a database connection so it could look things up, and now it technically has the power to delete every row in it.
IntermediateRead 6 minDoing: an hour or two6 steps
Decide between LangGraph, CrewAI, and a plain agent loop
Every roundup ranks a different framework first, and you just want to know which one fits the agent you're actually building.
IntermediateRead 6 minDoing: about an hour, mostly spent deciding rather than coding6 steps
Your agent wanders when a goal has no plan
You gave your agent a multi-step goal and watched it try something plausible-looking on step one instead of figuring out the whole path first.
IntermediateRead 7 minDoing: about an hour, working from a tool loop you already have running7 steps
Give a long-running agent a memory file it rereads every run
Your coding agent relearns the same project quirks and past decisions from scratch every single run.
IntermediateRead 6 minDoing: about an hour to build, then a few minutes at the end of every run after that7 steps
Run three independent LLM checks in parallel instead of one after another
Three separate checks on the same input run one after another, and the wait adds up every time.
IntermediateRead 5 minDoing: 30 to 45 minutes6 steps
Should your agent ask before it deletes, sends, or pays?
Your agent can already call the tool that issues a refund, and nobody decided whether it should do that on its own.
Intermediate to advancedRead 8 minDoing: a day7 steps
Draft a handoff contract between two agents
Two agents share one job, and you can't tell which one is supposed to answer a given message until you've read both of their prompts.
Intermediate to advancedRead 9 minDoing: an afternoon, mostly spent writing and testing the rule rather than coding7 steps
Stop a few big tool results from filling your agent's whole context window
Three tool calls into a long agent run, half your context budget is already gone and nothing in it is helping anymore.
Intermediate to advancedRead 10 minDoing: an afternoon, most of it spent tracing your own agent's tool calls and checking the token counts8 steps
Your agent never checks its own answer before handing it back
The output looks fine. There's no way to know if your agent is quietly wrong.
AdvancedRead 9 minDoing: a day8 steps

Cookbook badge

Finish all 14 recipes to earn this.

AI Agents cookbook

Write a tool-using agent loop that knows when to stop

One bad tool call crashes the whole agent run

Log every tool call so you can find what broke a run

Split an overloaded prompt into a checked three-step chain

Give a runaway agent a spending cap

Shrink an agent's blast radius by scoping down its tools

Decide between LangGraph, CrewAI, and a plain agent loop

Your agent wanders when a goal has no plan

Give a long-running agent a memory file it rereads every run

Run three independent LLM checks in parallel instead of one after another

Should your agent ask before it deletes, sends, or pays?

Draft a handoff contract between two agents

Stop a few big tool results from filling your agent's whole context window

Your agent never checks its own answer before handing it back

In these roadmaps

AI Agents cookbook

Write a tool-using agent loop that knows when to stop

One bad tool call crashes the whole agent run

Log every tool call so you can find what broke a run

Split an overloaded prompt into a checked three-step chain

Give a runaway agent a spending cap

Shrink an agent's blast radius by scoping down its tools

Decide between LangGraph, CrewAI, and a plain agent loop

Your agent wanders when a goal has no plan

Give a long-running agent a memory file it rereads every run

Run three independent LLM checks in parallel instead of one after another

Should your agent ask before it deletes, sends, or pays?

Draft a handoff contract between two agents

Stop a few big tool results from filling your agent's whole context window

Your agent never checks its own answer before handing it back

In these roadmaps