Make your codebase ready for AI agents.
An AI can build you an app. But when it changes something, will your project catch what it breaks? Have an AI examine the checks, tests and instructions around your code, show you where it still depends on a human spotting mistakes, and help you improve it. Try this on something you've built yourself or a project your team maintains.
From Feedforward's Praxis. Created by Joel Confino, Vanguard.
Use it on Praxis
This move runs as a ready-made skill in Praxis, Feedforward's skill library.
How it goes
What the AI will do
- Examine eight parts of your project, from tests that catch broken behavior to the instructions an AI follows when changing code.
- Give each part a score, show the evidence behind it, and explain which improvements would help most.
- Make the improvements you approve, then check that they work.
What to have ready
- Access to a software project: an app you've made, a repository your team maintains, or a copy of its files.
- Anything you've noticed going wrong when AI changes it. A rough description is enough.
Where you come in
- Explain what the project does and what you want AI to help with.
- Review the findings in plain language and choose which changes, if any, to make.
- Try the affected behavior after the changes and say whether it still does what you need.
What you'll get
A scorecard showing how well your project catches AI mistakes, a prioritized list of improvements, and—if you choose to proceed—changes that have been made and checked.
You're done when
A scorecard showing how well your project catches AI mistakes, a prioritized list of improvements, and—if you choose to proceed—changes that have been made and checked.
*Praxis is available to Feedforward members.