Skip to main content
Skip to content
Menu

Codex Beginner 3 parts Updated

Build a persistent context script

You'll have a working script that saves your coding project's decisions, past solutions, and preferences to a single file that you feed to AI on every request.

What you'll make

Project Context Script

  1. save_context() — captures project state and decisions: what to include
  2. load_context() — formats context for AI consumption: what to include
  3. update_context() — handles new decisions and constraint changes: what to include
  4. main() — orchestration and testing functions: what to include

Paste your working Python or shell script that you've tested on your real project

The 3 parts

Each part ends with a check you can see.

  1. Audit your project memory gaps

    Find the 3 decisions you most often re-explain to AI when starting coding sessions

    Done when

    You have a ranked list of 3 specific project decisions with examples of how often you re-explain each one

  2. Test AI with different context formats

    Try 3 different ways of presenting your top decision to AI and find which format it actually uses

    Done when

    You have 3 tested context formats with scored results showing which format made AI's output most immediately useful

  3. Build the update workflow

    Code the functions that save new decisions and test whether your context file stays current over one week

    Done when

    You have working save_context() and update_context() functions tested with at least 3 new decisions from real coding sessions

    Capstone: you put together your Project Context Script.

Where this helps

An elective: a single lesson that sits outside rtwo's job certificates and doesn't count toward one.

About this lesson

What you'll be able to do

  • Build and test a working persistent context script on your actual coding project
  • Identify what information AI needs to preserve vs what it can rediscover each session
  • Have a tested update workflow that keeps your context file current without manual overhead
  • Know when your context script is helping vs hurting AI performance on your real tasks

Who it's for

  • Developers who've used AI for coding but get frustrated re-explaining project decisions every session
  • Anyone working on a multi-week project who wants AI to remember past solutions and constraints

What it won't cover

  • Setting up specific AI tools or managing API keys
  • Advanced prompt engineering or fine-tuning techniques
  • Team-wide context sharing or collaborative workflows
  • Integration with specific IDEs beyond basic file reading

Skills you'll practise

  • Context file design
  • Information persistence testing
  • AI memory validation
  • Context update workflows