Learn prompt crafting by building useful writing tools

Your starting point

You can describe a small feature in plain language. These projects can run locally without calling an AI service.

• Basic form inputs and text output
• A small example project idea and a few realistic scenarios

Your finish line

You can write observable requirements, compare outputs with a stable rubric, and hand off current facts without inventing verification.

Stack up four small wins

1. STEP 1 OF 4

   Rewrite one vague request

   Rewrite “make a better task app” as a single useful journey and four checks.

   Bro tip: Underline the user action and the visible result in your request. Replace vague praise words with checks a reader can actually apply.

   What you’ll learn

   • Goal versus preference
   • Current versus desired behavior
   • Observable acceptance criteria
   You’re ready to move on when

   A reader can tell whether the requested work is complete from the stated behavior.

   I can do this step

2. STEP 2 OF 4

   Build a reusable brief composer

   Build a form that assembles a clear build request from the user’s own answers.

   Bro tip: Leave a required field empty on purpose and inspect the preview. It should show what is missing instead of inventing project facts.

   What you’ll learn

   • Relevant context
   • Scope boundaries
   • Copyable output
   You’re ready to move on when

   The preview includes the entered requirements and never fabricates missing project facts.

   Open project: Build brief composer

   I can do this step

3. STEP 3 OF 4

   Evaluate a prompt revision

   Build a small prompt experiment notebook with manually recorded outputs.

   Bro tip: Keep one required case visible while comparing totals. Read the actual output for that case before calling a prompt revision better.

   What you’ll learn

   • Representative cases
   • Scoring rubrics
   • Regression comparison
   You’re ready to move on when

   Candidates use the same cases and rubric, and a better average cannot hide a critical regression.

   Open project: Prompt experiment notebook

   I can do this step

4. STEP 4 OF 4

   Carry forward the current state

   Build a handoff composer that keeps goals, decisions, checks, failures, and the next task distinct.

   Bro tip: Read the packet as if you missed the whole conversation. Find the next task and distinguish observed results from planned checks.

   What you’ll learn

   • Context packets
   • Observed versus planned work
   • Verification evidence
   You’re ready to move on when

   A fresh reader can continue from the packet and identify which results were actually observed.

   Open project: Context and handoff kit

   I can do this step

Your build lineup

1. Build brief composer

   Turn a small app idea into a clear, copyable request using the user’s own requirements.

2. Prompt experiment notebook

   Record prompt versions and compare manually entered outputs using the same cases and rubric.

3. Context and handoff kit

   Build a compact project packet that separates current facts, plans, assumptions, and verification evidence.

Watch it. Then build it.

Start with a worked example, then connect the recording to your own small, checkable build. Videos open on their creator’s or publisher’s website.

Read Vibe Bro’s written lesson briefing ↗

1. Beginner · Video · 23 min

   Understanding Prompt Engineering Fundamentals (Part 4 of 18)

   Microsoft Learn · Nitya Narasimhan

   An introduction to prompt components, iterative revisions, and checking whether a response meets the task.

   For step 1: Rewrite one vague request

   Before watching: No machine-learning background required. Reading basic code helps with the optional API example.

   Watch, then try

   Rewrite one vague build request as you watch. Save its goal, relevant context, and one observable completion check in the brief composer.

   The lesson demonstrates an OpenAI endpoint. Watching and doing our brief-composer project do not require an API account or a paid model call.

   Watch video ↗ · Practice: Build brief composer

   Publisher’s course details · checked October 28, 2026

2. Next step · Video · 17 min

   Creating Advanced Prompts (Part 5 of 18)

   Microsoft Learn · Chris Noring

   A follow-up lesson on changing a prompt deliberately and comparing the outputs produced by different instructions.

   For step 3: Evaluate a prompt revision

   Before watching: Watch the fundamentals lesson first; basic familiarity with model inputs and outputs helps.

   Watch, then try

   Pick one prompt change and compare saved outputs on the same examples. Record the change and a concrete success or failure in the experiment notebook.

   This is a more technical follow-up. Apply the ideas to saved examples first; model names and service setup shown in older demonstrations may differ from your current tools.

   Watch video ↗ · Practice: Prompt experiment notebook

   Publisher’s course details · checked October 28, 2026

Video links selected October 28, 2026. Selection checks cover publisher metadata and topic fit.

The docs behind it

These references explain the underlying concepts and APIs. The prompt templates and project plans are original examples to adapt to your own setup.

• OpenAI — prompt instructions, examples, and context

Documentation links checked October 27, 2026.

Keep the momentum going

• Abstention policyCoding term
• Add one featureCopyable prompt
• AgentCoding term