There’s a common pattern of building up a council of reviewers and have them debate a topic. The first ones of these I saw was the Council of HAL. Following that, I’ve seen similar things like with open-code-review, which will bring up a panel of experts and have them debate, then synthesize the feedback and write it into a PR.

In doing some design doc review, I stumbled on a pattern I think is really interesting: asking the robot for a recommendation of other colleagues who might have good insights.

You’re a principal engineer, responsible for signing off on this implementation plan. You have expertise in large scale migrations. You understand the practicalities of the real world. Be exhaustive, accurate, honest and helpful. Review the proposal as written. Refer three of your other colleagues you think should take a look, as well as a brief paragraph describing their expertise/perspective and what you’d hope to get from their review.

followed by :

Then, spawn a subagent for each of the personas suggested. Tell them to conduct similar analysis. They can look around in the folders in ../ which should be all of the potential affected systems. Have them each write their own findings documents

One of the things I thought was interesting is that after synthesizing this feedback into the document and running this a second time, different experts were chosen, presumably because the original concerns were addressed.