The 7 Prompting Habits of Highly Effective Engineers *

2025-05-20 by Josh Bleecher Snyder

* I would have preferred to title this 13 Ways Of Prompting an Agent, but apparently more people know about Steven Covey than Wallace Stevens.

In the 2000s, we learned to Google.

In the 2010s, we learned to ask Stack Overflow.

In the early 2020s, we learned to prompt LLMs.

That was so last year. Now that it’s 2025, we’re learning to prompt AI Agents, such as sketch.dev (us), Claude Code, OpenAI Codex, and oh so many more. (Hi, friends!)

This is post is a list of techniques that we’ve found to be effective for working with coding agents.

Draw the Owl

Make a code change illustrating what you’re up to. Ask the agent to finish the job. Or do a thing all the way once and tell the agent to read that commit and repeat it.

Meme titled "How to draw an owl." Left panel: two rough overlapping circles labeled "1." Right panel: a fully detailed pencil sketch of an owl on a branch labeled "2." Captions below read "1. Draw some circles" and "2. Draw the rest of the beautiful owl," with "beautiful" scrawled in blue marker.

Use a checklist

Four-panel Gru presentation meme from Despicable Me. Panel 1: Gru points at a blank flip-chart. Panel 2: Chart now reads "Ask agent to do tasks A, B, C, D." Panel 3: Chart reads "A: done!" and Gru grins proudly. Panel 4: Chart reads "B: squirrel!" and Gru looks confused.

For bigger projects:

  • Ask the agent to make a plan and break it down into a detailed checklist.
  • Write the checklist (temporarily) to your repository.
  • Have it work through the checklist, checking items off as it goes.
  • Delete the checklist.

Send out a scout

Hand the AI agent a task just to find out where the sticky bits are, so you don’t have to make those mistakes.

Disaster-girl meme: a young girl smirks at the camera in the foreground while a house burns behind her. Top text: "THIS BUG LOOKS EASY." Bottom text: "OH HUH."

Fail early, fail often

IQ bell-curve Wojak meme. A normal-distribution graph fills the center; left "low IQ" Wojak says "Make logging in work," the center "mid IQ" Wojak gives a long, detailed bug report above the curve, and the right "high IQ" hooded Wojak repeats "Make logging in work."

Agents are pretty good at getting oriented, and failure is cheap. Throw out your problem, and let it work at it. If it fails, start over, adding the tidbits of information it couldn’t figure out on its own. If the agent goes off on a tangent, don’t try to redirect it or fix it. Cut bait and start over with a similar prompt.

Go Concurrent

Agents tend to be slower than normal human attention span. You can hide that latency by firing up a bunch of them and then checking in with them when you’re ready. You can also have multiple agents tackle the same task and compare the results.

Highway-swerve meme. Upper image: freeway sign—the left lane arrow says "Start one agent," the right exit arrow says "Start ten agents." Lower image: a car screeches across the solid lines to take the exit.

Provide goals, not instructions

Business-cat meme: a black cat wearing a neck-tie sits against a colorful radial background. Top text: "YOU CATCH THE BUG." Bottom text: "I’LL TAKE A NAP."

Whether it’s speeding up your tests or fixing a class of eslint warnings or iterating on browser screenshots, if the agent can evaluate success, let it do its thing.

Farm out step 0

There are a variety of “deep research” products out there. Have them prepare detailed docs on an API or technology, and pass that along. There are a bunch of UI prototyping tools out there. Explore there and then let the agent integrate.

Meme of a sloth reclining lazily on a tree branch in a jungle. Top text: "YOU FINISHED STEP 1?" Bottom text: "YAAWWN. ON TO STEP 2, THEN."

sketch.dev · merde.ai · pi.dev