I use OMP. Was on Pi but switched over because I like OMP's subagents implementation. Use 5.5/4.8 for coming up with a very detailed plan, ds4pro/mimo for implementation of easy features, 5.5/4.6 for larger/harder features, and 5.5/4.6 for a review. I use them in onorca.dev(settled here after trying just about any "agent ade". i also sometimes use Zed as well. I'm curious to know if folks use any meta-harnesses?
Work I have access to both Claude Code and Codex. Use them both, typically on the same work. I like to do brainstorming and solution design with Codex GPT 5.5 High, and then have Claude Opus 4.7/4.8 high/medium do the implementation.
So far been quite productive for some greenfield rewrites and refactors of existing code. Though bug fixing our main project it's more hit and miss, though Codex GPT 5.5 High can be very good at spotting subtle issues.
How do you a) share context/chat etc with each other? plugin? have them make a file? Why did you opt for either of them instead of something like opencode that supports those models?
Currently we're mostly working on separate issues or aspects, so not much context sharing needed. We have some skills that we share in a shared network directory.
As for why using both instead of OpenCode or similar, well, subscriptions take you a long way for not very much. Also security is a thing here now so sec guys feel more comfortable with big names rather than "random" open source.
I have been building my own coding agent, VT Code (https://github.com/vinhnx/vtcode). I have now made it stable and mature enough, and I use it for most of my works.
Mimo Code has grown on me. Before that, a very janky Pi. I've realized that at one point I was spending too much time getting things exactly as I want them versus actually using the damn thing
At work we only have access to claude / kiro. I only use it via cli since I find this to be the most flexible setup as claude can find stuff anywhere, read, write and run programs etc.
claude code + codex，for sure. I also actively try Other runtimes.
I use OMP. Was on Pi but switched over because I like OMP's subagents implementation. Use 5.5/4.8 for coming up with a very detailed plan, ds4pro/mimo for implementation of easy features, 5.5/4.6 for larger/harder features, and 5.5/4.6 for a review. I use them in onorca.dev(settled here after trying just about any "agent ade". i also sometimes use Zed as well. I'm curious to know if folks use any meta-harnesses?
Whatever external (none-local) provider you use make sure to have backup plans. My horror story is below;
https://news.ycombinator.com/item?id=48597861
I use a mix of claude code, codex and cursor (composer 2.5). Whichever has more credits left.
Work I have access to both Claude Code and Codex. Use them both, typically on the same work. I like to do brainstorming and solution design with Codex GPT 5.5 High, and then have Claude Opus 4.7/4.8 high/medium do the implementation.
So far been quite productive for some greenfield rewrites and refactors of existing code. Though bug fixing our main project it's more hit and miss, though Codex GPT 5.5 High can be very good at spotting subtle issues.
How do you a) share context/chat etc with each other? plugin? have them make a file? Why did you opt for either of them instead of something like opencode that supports those models?
Currently we're mostly working on separate issues or aspects, so not much context sharing needed. We have some skills that we share in a shared network directory.
As for why using both instead of OpenCode or similar, well, subscriptions take you a long way for not very much. Also security is a thing here now so sec guys feel more comfortable with big names rather than "random" open source.
Opencode and other harnesses like Pi do support codex and claude(for now) subscriptions, but the point about security makes sense.
I have been building my own coding agent, VT Code (https://github.com/vinhnx/vtcode). I have now made it stable and mature enough, and I use it for most of my works.
Ohh nice! Maybe I'll try it, in your opinion whats the best/defining feature?
Mimo Code has grown on me. Before that, a very janky Pi. I've realized that at one point I was spending too much time getting things exactly as I want them versus actually using the damn thing
I really like Grok Build. Its TUI interface, sleek design, and ease of use have really won me over.
I use Kilocode Vscode extension, still can't give up the IDE
At work we only have access to claude / kiro. I only use it via cli since I find this to be the most flexible setup as claude can find stuff anywhere, read, write and run programs etc.
Both, depends on which has the best model and/or the task I am doing. Frontend always CC.
I use my brain, pen and paper, because I work on serious projects that are of high risk and are mission critical. Not toy projects on GitHub.
Secondly, there will be a time where companies like Anthropic will disallow you from using AI in their interviews.
So you are cooked if you are over-reliant on coding agents.
A.ctual I.ntelligence
Mostly just Codex
claude - org pays for it. org used to pay for both, and i preferred codex back then (~3mo ago).
i find claude to be significantly lazier or require significantly more guidance. it does, however, have better design and ui/ux intuition.