Hey HN,

I am a professional developer who loves building side projects and early adopting AI.

After I tried OpenAI Codex, I was impressed. I can write a prompt and get a PR.

When Anthropic released Claude Code SDK, I started building AgentWrap to:

1. ship from my mobile phone, 2. ship multiples features in parallel.

I personally use and improve the service every day and hope you find it useful too. Would love to hear your feedback and suggestions.

Denis

Sounds like something that's just one release announcement away from being built by the platforms as it's already exactly what https://docs.github.com/en/copilot/how-tos/use-copilot-agent... does.

Claude can already do this: https://docs.anthropic.com/en/docs/claude-code/github-action...

Nice, but there are self hosted / open source systems that allow you to do the same ; here I have to give my access to someone I don't know who might also use my claude after analysing usage and seeing I have plenty of credits left or when I sleep? Not a bad idea actually but :)

I'm not saying you are, but how can you prevent it? I don't think we can see usage of the plan ones (API of course we can) right? I'm not sure how anyone can guarantee?

how does it handle prompts for user input?