We have

good first issues

but we don't have labels beyond that to let contributors know the next level issues. Would y'all want a

good second issue

label? This came up because I have talked with @nice-airplane-52666 and @bright-table-36639 who both want to continue to contribute and find good issues. But after

good first issues

it becomes less clear what the next issue should be tackled. The weight of the next issues is unclear. So looking for feedback in anyway we can help you contributor more 🙌

Just a few thoughts You can always use OpenHands-CLI to do the stuff you’d like, and see what happens. See what you don’t like, what you’re missing, what breaks. Then tell it to fix that. Another thing is, send the agent on an issue, but not to solve it immediately, to investigate it instead. To understand it, and to help you understand how it could be and its scope. Maybe break it down. Maybe send the agent to only write real-like tests that should fail as long as the issue isn’t solved, first, and then tell it to fix it. Maybe you could tell the agent to analyze OpenHands V1 execution paths, and try to complete the docs in docs/ repo. I don’t know, just some thoughts. I do want to note though: SDK and CLI repos are much smaller than the Web repo, relatively well contained and ready for agent use (we hope), so that the agent doesn’t get fooled into weird paths that exist in the legacy codebase of the web repository. So it should be easier to play with - again, hopefully 🙏

+1 to

good second issue

label this is exactly the gap I've been hitting after clearing good first issues I've been doing a lot of reading through the codebase to find the next thing, which works but it's slow. Even a rough difficulty/scope label (e.g. medium complexity, requires understanding of X subsystem) would help a lot. Happy to help dogfood whatever system you put in place.