The Adversarial Review System Every AI Coding Agent Needs with Wes McKinney at Kenn Software

Was that thinking the genesis of RoboRov? It was.

It took a little while to come to the realization.

Like I really enjoyed like just the general vibe of using cloud code to borrow a term.

But then I found that cloud was simply not that good at doing code review.

And In the meantime, OpenAI shipped their first coding agent, Codex, and I didn't really enjoy the experience of using Codex to build software.

Just it felt a little bit slow, like the agent a little bit deliberative, like not very eager to do things, a bit timid.

But what I found just by experimentation was that Codex and GPT-5 in particular was tremendously good at finding bugs in code, like finding edge cases, issues in my test suite, finding security issues, things like that.

And so essentially what I found myself doing was running Cloud Code in one terminal pane and then Codex in another terminal pane.

And anytime Cloud would do something, I would have it commit immediately.

And then in my Codex session, I would say, okay, Codex, review this commit.

And then after a while, I was doing that so many times a day.

I was like, why don't I just automate this? Like, why do I even have to ask? It should just happen automatically.

And so that was the first week of January this year.

Like the first iteration of RoboRev was just a very simple daemon that would be triggered by a commit hook, post-commit hook.

So whenever... you tell your Cloud Code sessions in Cloud.md, every time I prompt, like every time you return control to me, never leave the Git repository in a dirty state, like never leave uncommitted changes, always commit.

And so that creates like the commit loop.

So anytime the agent does anything, it commits.

And that triggers a code review that runs in the background.

And then initially, I was just copy and pasting the code reviews from the RoboRev system into Cloud Code, which got me pretty far.

But eventually...

It became more hook based where you ask, especially during like marathon implementation sessions, especially like using some of the spec driven development frameworks.

You want it to automatically pull in the code reviews that have piled up while it's implementing.

So I'm a superpowers user, for example.

And so I might implement like a 40 task superpowers plan and then I'll tell.

The superpowers implementer harness run RoboRef fix every five tasks to address code review feedback that is piled up in the background.

Was that thinking the genesis of RoboRov? It was.

It took a little while to come to the realization.

Like I really enjoyed like just the general vibe of using cloud code to borrow a term.

But then I found that cloud was simply not that good at doing code review.

And In the meantime, OpenAI shipped their first coding agent, Codex, and I didn't really enjoy the experience of using Codex to build software.

Just it felt a little bit slow, like the agent a little bit deliberative, like not very eager to do things, a bit timid.

But what I found just by experimentation was that Codex and GPT-5 in particular was tremendously good at finding bugs in code, like finding edge cases, issues in my test suite, finding security issues, things like that.

And so essentially what I found myself doing was running Cloud Code in one terminal pane and then Codex in another terminal pane.

And anytime Cloud would do something, I would have it commit immediately.

And then in my Codex session, I would say, okay, Codex, review this commit.

And then after a while, I was doing that so many times a day.

I was like, why don't I just automate this? Like, why do I even have to ask? It should just happen automatically.

And so that was the first week of January this year.

Like the first iteration of RoboRev was just a very simple daemon that would be triggered by a commit hook, post-commit hook.

So whenever... you tell your Cloud Code sessions in Cloud.md, every time I prompt, like every time you return control to me, never leave the Git repository in a dirty state, like never leave uncommitted changes, always commit.

And so that creates like the commit loop.

So anytime the agent does anything, it commits.

And that triggers a code review that runs in the background.

And then initially, I was just copy and pasting the code reviews from the RoboRev system into Cloud Code, which got me pretty far.

But eventually...

It became more hook based where you ask, especially during like marathon implementation sessions, especially like using some of the spec driven development frameworks.

You want it to automatically pull in the code reviews that have piled up while it's implementing.

So I'm a superpowers user, for example.

And so I might implement like a 40 task superpowers plan and then I'll tell.

The superpowers implementer harness run RoboRef fix every five tasks to address code review feedback that is piled up in the background.