Francesco

I want to grow and market on X, but I'd rather focus on building the business I want to post more, but it takes up a lot of time. I feel like I should use AI to write faster, but nobody enjoys reading AI slop I solved this and open-sourced it. Here's the sauce 👇

How to run virtually infinite app instances without localhost port conflicts

Critical SSH vulnerability in Alpine Linux lets you log in as root with any password longer than 257 characters yes, any password. yes. I know. yes. Requirements: - sshd compiled with musl - PAM not being used - password logins enabled - at least one account with "*" as the password hash - at least one other account with a valid password All of these are extremely common in the default Alpine configuration

I shipped 71 PRs in the last three days, just on side projects The unlock was letting agents work in parallel without merge conflicts, port conflicts or shared accounts Codex turns ideas into tickets, then each agent works unattended on a ticket in a separate git worktree Vercel's portless library gives each worktree its own app instance on its own .localhost URL, so agents can test in the browser without fighting over ports Vercel's Emulate gives each worktree its own local, stateful mocks for Stripe, Google OAuth, Gmail, Resend, and GitHub, so there are no shared sandbox accounts or per-instance OAuth and webhook setup Without isolated worktrees, app URLs and integrations, parallel agents just step on each other. With them, I can run 10+ instances at once and let the work continue unattended for entire days The agentic harness that packs all of this is open-source on GitHub

I run and test 10+ local instances of my app in parallel without conflicts, and now each one gets its own mocked Stripe and Google server. My agents won't get stuck on third-party API problems ever again My open-source agentic harness uses git worktrees and Vercel's portless library to run virtually unlimited local instances without port or env conflicts I added Vercel's Emulate library too. It runs local, stateful mocks of the third-party services my apps depend on: Stripe, Google OAuth, Gmail, Calendar, Drive, Resend, Clerk, AWS, GitHub, Twilio, and more Every worktree gets its own fake Stripe and fake Google, so there's no shared sandbox account, and no need to change the OAuth redirect URLs, webhooks, and rate limits for every running instance It's open source, go steal it and enter the 10x dev mode

How to get ahead of 99% of people: - sleep 8-9 hours - have a purpose to pursue daily - lift weights 3x week - eat 1.5g protein per kg of body weight a day - 5g creatine a day - cut toxic relationships

THAT'S how I want to see outages handled I'm calling out indie hackers here. Even with a small audience, the way you handle problems for your users matters. It's about trust, and letting them see that you're there fixing the product they're paying for, and you're not leaving until it's back

I solved the localhost port conflicts problem forever Vercel's portless library lets my agents test virtually infinite app instances in parallel, without conflicts I added it to my open-source agentic harness sdlc.pumpkit.io Everything runs on autopilot with zero conflicts Now my agentic harness has no limits other than my speed of putting my ideas into Claude Code I use Claude's voice mode a lot already, but it's still not enough It looks like I'll have to build an orchestration layer to make more decisions for me Steal the repo and start shipping 10x

Absolute banger

I ship 1000+ PRs/mo, and it doesn't happen by mistake I built an agentic harness to keep agents running, and open-sourced it Here's the flow:

Should you feel bad about using AI to write your posts? If you ask me, no. Here's why. Since we started using AI, we've used it for everything. We even outsourced the creative part to it, then got lazy and started expecting it to do the repetitive work and the thinking too, while somehow producing magical results You can usually spot AI-written content on social media immediately. The same phrases, structure, rhythm and generic slop show up over and over again But those tells are fixable The useful part of AI is automating repetitive work. In this case, the repetitive work is writing the post after you've already had the ideas If an AI system can write in your voice, remove the slop and turn your ideas into something people actually enjoy reading, I think it's fine to use it You're saving time with a tool. The content is still yours. The ideas and thinking are yours That's what I'm also doing. I open-sourced the web app I built to write X posts, remove AI-slop, and schedule the post for me, so that my remaining duty is the creative thinking It also has an MCP that Claude Code and Codex can use - I rarely use the web app myself All open-source. Link to the GitHub repo in the replies

I ship 1000+ PRs a month, most of them while I sleep My job is mostly agreeing with the agent on what to build Specs, tickets, code, review and PRs happen automatically The whole harness is open source