10 open-source repos worth knowing if you build with coding agents or post about it on X. All updated this month.
For building:
obra/superpowers: a skills framework plus a dev methodology for your coding agent
https://t.co/C4f8ZDbUda
cursor/plugins (pstack): verify-first skills behind 4,000 merged PRs in a month, by its author's count. Installs in Claude Code too. Its skills have been on my machine since last week
https://t.co/ADHhhpAzSN
upstash/context7: current library docs for your agent, so it stops coding against old APIs
https://t.co/kf9rDriWyS
SWE-agent/mini-swe-agent: a ~100-line agent. A new paper found it within 5 points of Claude Code on SWE-bench Verified (https://t.co/vK58FbU2NK), with an 829-token preamble vs Claude Code's 16,581
https://t.co/1D0r2kIvY2
gitleaks/gitleaks: catches API keys before your agent commits them
https://t.co/nuBfddHUBk
Nutlope/hallmark: a design skill that keeps agent-built UIs from all looking the same
https://t.co/aPWP7SOgNC
For posting on X:
xai-org/x-algorithm: the code behind the For You feed. Read what it rewards before you write
https://t.co/zboWVWumMs
gitroomhq/postiz-app: open-source scheduler for 28+ channels including X, built for agents
https://t.co/7hryI2wq79
remotion-dev/remotion: make videos in React, so Shorts and clips come from code
https://t.co/6ULTL4dX8s
excalidraw/excalidraw: hand-drawn diagrams that stay readable in a feed
https://t.co/WMgni177rG
Start with x-algorithm. It's the one that shows how your posts get ranked, and it changed how I write mine.
Introducing e-means, an algorithm beating k-means in 10x fewer epochs.
With order of magnitude wall-clock speedups, e-means is released in FAISS and needs no hyperparameter tuning beyond k. It is robust to various scales and datasets. Still skeptical ? Just try e-means [1/10]🧵
@GoogleColab is there any on-going issues with your GPU runtimes ?
I tried multiple times but it failed all the time, and I still have 100 units of compute left.
Please help me with this.
@oimatrixioYT@marclou I have few things in my mind a big problem and few smaller problems, I will open source the solution to all smaller problems in coming days and parallely I will work on solving the bigger problem.
Marc Lou’s rule is “don’t build if I’m second.” I’m new to solo dev, and I’m breaking it on purpose.
I pick an app on @marclou’s TrustMRR, where the revenue is verified, and build a cheaper version of it. So far that’s Roastometer, which roasts your homepage, and Looplight, which makes looping animated backgrounds. Both are live.
I’m a little embarrassed the ideas aren’t mine. I’m also proud. Two apps are live that real people can use, and I built them next to a 9-5. And I still find it hilarious that people pay for apps like these.
To the founders who built the originals, I’m learning from you, not coming for your customers.
Is being second is fine while learning things,
now I'm thinking of solving my own problems (excited to see where it goes)
Hey paul not really I'm thinking of solving my own problem, i got the taste of what it takes to build and launch a product online for others to use.
And still my two apps work cheaper and best, I don't find the energy to do marketing, I think I need to put in more thoughts on solving my own problems and bring solutions, may be that way I find more meaningfulness and energy in my products.
Nice to meet you Paul !
@DanielSmidstrup@grok Hmm in general, the chances of getting a positive response from being good are higher than expecting positive results from negative actions.
Google 9. OpenAI 7. Anthropic 5, with the verdict “Four green dots in a petri dish, and Claude only lives in the menu.”
Your first roast is free now. You get a score, three jokes about your actual page, and the one fix worth making.
https://t.co/csuzaHyJ25
Five Elon Musk companies, and not one homepage scores above 6 out of 10.
Tesla 6. SpaceX 6. Neuralink 5. X 5. The Boring Company 4, with the verdict "Raised $3B, and none of it went on a button."
Google got 9.
I built https://t.co/IAB5rai5sT to grade landing pages like these. Paste yours and you get a score, three jokes about your actual page, and the one fix worth making, for $1.
Five Elon Musk companies, and not one homepage scores above 6 out of 10.
Tesla 6. SpaceX 6. Neuralink 5. X 5. The Boring Company 4, with the verdict "Raised $3B, and none of it went on a button."
Google got 9.
I built https://t.co/IAB5rai5sT to grade landing pages like these. Paste yours and you get a score, three jokes about your actual page, and the one fix worth making, for $1.
Paul, I don’t want to label anything, but I’d like to explore all parts of tech. I grasp the why and what as soon as I study something, so I have no difficulty understanding a technology. Yet I’ve failed to stay consistent in certain things that don’t make sense one example is solving coding problems(leetcode style). Building a product lets me sit for hours thinking and crafting it, but solving coding problems is just dopamine right now.
Nice to meet you !!