Salbotics Solutions

Salbotics AI
Harness Lab

Beginner-friendly talks, worksheets, and templates for turning AI experiments into repeatable workflows without losing the joy of exploration.

Start Here

Harness engineering is a practical structure around AI work: clear goal, useful context, visible path, review checkpoints, and a definition of done.

Goal + Context + Path + Checkpoints + Done Criteria

For never-coders

Use the worksheet to turn one repeated task into a prompt that stays understandable.

Open worksheet

For vibe coders

Add checkpoints so experiments do not turn into mystery files and endless fixes.

Copy checkpoint rule

For teams

Package domain terms, examples, prompts, and review gates into a small context pack.

View context pack

Episode 1

Implement Harness Engineering: A Clear Path to Minimize Rabbit Holes While Vibing.

Templates

Copy one template, use it three times, then improve it once. That is enough to start a prompt library.

Harness Prompt

Goal, background, inputs, constraints, path, checkpoints, done.

Open template

Context Pack

A small Markdown folder for one repeated workflow.

Open starter

Checkpoint Rule

The cheapest way to stop drift before it gets expensive.

Open rule

Episode 2

Lessons From Real Projects is a completed public-safe tutorial pack for non-technical builders. It teaches work orders, context packs, checkpoints, testing, and redaction using fake examples only.

AI Work Orders

Turn a vague idea into a reviewable prompt brief.

Open tutorial

Privacy First

Share useful lessons without exposing private project details.

Open safety map

Worksheets

Use prompt, checkpoint, redaction, and handoff checklists.

View episode

Episode 2 Checkpoints

The tutorial pack is finished through CP5. Each checkpoint leaves a master list so future updates do not need to rediscover the context.

Episode 3: Programming Language SWOT

A student-focused module for understanding what programming languages are used for, how each one benefits a learner, and how to compare strengths, weaknesses, opportunities, and threats over the next 5-10 years.

Language Applications

Group languages by function: AI, web, data, backend, systems, mobile, research, games, blockchain, and legacy work.

Open module

Full SWOT Report

Review the top 25 career-relevant languages with student benefit, applications, outlook, and SWOT.

Read full report

Exercise Next

The next checkpoint will turn this topic into practical classroom activities so learners can test language-function understanding.

View placeholder

Current Labs

Public-safe game labs show how checkpoints turn ambitious ideas into playable, testable demos without losing scope control, source attribution, or handoff notes.

Lab 02: Ular

A public-safe snake roguelite lab through Pyxel CP11 with modules, hazards, boss pressure, scrap economy, and developer mode.

Open playtest tutorial

Lab 03: Underdog

A Phaser + TypeScript racing lab with simcade handling, Garage upgrades, rival duels, championship loop, difficulty, and visible steering feedback.

View racing lab

Underdog Tutorials

Learn the working wisdoms: harness big ideas, build racing feel, tune fair pressure, and test with evidence.

Start tutorial path

What Underdog Teaches

The racing exercise is a practical study in making ambition usable. It shows how to move from idea, to scaffold, to physics, to upgrades, to rivals, to championship pressure, while keeping each step testable.

Feel Before Features

The car had to explain steering, speed, surface, and mistakes before the campaign systems could matter.

Read the lesson

Fair Pressure

Difficulty became a visible, saved, testable choice instead of a hidden punishment.

Read the lesson

Wisdoms

Big ideas need small doors. Tests need screenshots. Handoffs are part of the product.

Read the wisdoms