They seem to want to go broad with the tooling. At least this is the impression I got when reading their WWW. Nothing wrong with it. I suppose eventually they would have 5 to 10 more things than just the Merge Queue and the Flaky Bot.
I'd be curious what other ideas people might have for tools like this.
Let me start with a few from myself. Curious what others say:
1. A better Dependabot, which at least in Java (Maven/Gradle) ecosystem works only for basic projects. It has so many limitations.
2. Onboarding helper. I.e. something which sets up everything for a new developer, and maintains that consistency of the setup going forward. So that the infra/DevOps teams (who I guess would be the key users at your customers) can use it to provide consistent but flexible setup for everyone. And be able to evolve that. I mean things like IDE, build script setup with all the dependencies, key tools installed like terraform, GH client, a linter, etc. All of them in the same version for everyone.
3. Command-line friendly and user friendly credentials interface. Think of a mix of 1Password and hashicorp vault. But it should work well in all three contexts: nice GUI, CLI and automation (not only on my laptop, but as part of some CI pipeline or other toolling). It doesn't have to be the storage, delegate to 1P/HCV.
4. A collaboration tool for dealing with different kind of error messages, warnings and problems. Think of something like a mix of StackOverflow and Sentry. Group by type of the error, but focus on what developers experience in their dev environment (tests, build scripts, etc.). Deduplicate, store and offer solutions as a knowledge base
5. Time measurement tool, but not the one for my employer to check if I waste time on Facebook. But measure the length of my cycles - i.e. count of local runs before I commit, or count of CI-failure-then-fix iterations. And aggregate that, so that again my DevOps team can see some stats and know how to increase velocity by knowing what are the bottlenecks (but don't trust subjective feelings on that, gather data).
"Persistent project-oriented desktops"
Project-oriented means: I can have an entire desktop for say, work another for hobby, another for home usage. I don't want to mix windows/configurations between these contexts.
Persistent means: The desktop is like a physical room. It never "collapses" on its own, such as a during a restart, etc. In a physical room, if we place a chair somewhere, it doesn't move to a new position on its own. Similarly, given a room ("project desktop"), I don't want objects to move, unless I move them.
Despite having 100s of distributions, etc, there is no solution to this. There is neither persistence, nor good organization/segragation work into streams.
There might be market for such a tool for less-popular programming languages. For example a lot of tools overlook Perl (which cpan-minus and cpan-outdated the tooling is similar to other languages).