d-yoda

Anthropic would say that's exactly why they're in the red.

That's awesome!

How is the collateral handled? Is it like the agent borrows from platforms such as AAVE or Morpho?

I have never experienced a situation like this, so I can only assume there is a problem with the harness or something...

I think it would be better to have support for Skills.

I feel the pool is often way more fascinating than the actual front-pages nowadays.

So it’s basically similar to .onion on Tor Browser, but without requiring a blockchain. The idea sounds interesting. If you can find strong use cases for it, it could be very promising. (Monetizing OSS is always difficult.)

Isn't high quality defined by how easy it is to maintain as the scale grows? I feel there is a disconnect between the "quality as perceived by developers" and the "quality as experienced by designers and users."

I agree with most of this. I think code quality has been improving though. Compared to a year ago, the difference is pretty noticeable.

I ran into many of the same issues, and they motivated me to experiment with a linter that flags duplication and architectural problems across a codebase. It’s still a work in progress though:

https://github.com/ludo-technologies/pyscn

Hi HN. I built a CLI to install skills for AI coding agents from GitHub. Supports 30+ agents including Claude Code, Cursor, and Copilot.

Feedback welcome.

I recently reviewed some code written by a non-engineer using Cursor's vibe coding. It was full of dead code and patchwork logic. Vibe coding can build a working prototype, but maintaining it? That's another story.

So I built a GitHub App that audits your Python codebase. It reviews PRs and runs weekly scans on the entire repo, catching complex functions, dead code, and duplicates. When it finds issues, it posts comments and Issues with concrete suggestions.

It's powered by pyscn (https://github.com/ludo-technologies/pyscn), a static analyzer I open-sourced a few months ago.

Happy to answer questions.

I built a GitHub App that monitors your Python codebase for quality issues.

It reviews PRs and runs weekly audits on the entire repo. When it finds problems, it posts them as comments or Issues with concrete suggestions.

I open-sourced pyscn ([https://github.com/ludo-technologies/pyscn](https://github.com/ludo-technologies/pyscn)) a few months ago. I liked what it could do, but wanted to make it more accessible and more powerful. Combining it with an LLM felt like the right next step. pyscn-bot is the result.

It catches things like complex functions, dead code, and duplicated logic. It can also analyze your entire codebase architecture, something most review bots can't do since they only look at diffs.

Happy to answer questions.

So what happens when they swap Claude for Llama under the hood?

Thanks! I will fix the issue as soon as I can in the next version.

Yes, that's definitely something I'm looking into. I've received a few comments about AI coding integrations, so it's becoming a high priority.

This is super insightful, thank you for sharing. It's a great list of common LLM-generated anti-patterns.

I'd love to look into incorporating checks for these into pyscn. This is exactly the kind of stuff I want it to catch.

Wow, was it really in some newsletters? That's awesome to hear, and would definitely explain the recent spike on GitHub!

Thanks a lot for the bug report and for providing the details. I have a hunch—it's possible that you need to explicitly specify the path depending on your directory structure. For example, if your Python files are under a src directory, could you try running it like [your_tool_name] analyze src/?

If that still doesn't solve the problem, it would be a huge help if you could open a quick issue on GitHub for this.

Thanks again for your feedback!

Thanks for trying it out! For the progress bar bug, would you mind opening a GitHub issue with details? That'll help us track it down.

Great idea using it as grounding for AI-assisted refactoring! Let us know how that workflow goes.

Haha, maybe your code is better than you think! But seriously, we'd love to hear if the thresholds feel too lenient. Feedback welcome!

Thanks for sharing! Good to know about rust-code-analysis. Always helpful to see what other tools are doing in this space.

Fair point! My initial target was engineers using AI, but I'm open to refining the messaging.

MCP integration could be a good option. If there's interest, feel free to raise an issue on GitHub.

Pylint's duplication check is text-based (compares lines), while pyscn uses tree edit distance on ASTs. This means pyscn can catch structural clones even when variable/function names differ.

They complement each other - Ruff for style, pyscn for architecture. pyscn focuses on structural quality - checking if your code follows fundamental design principles like DRY, YAGNI, or other best practices.

Rust is definitely the king of performance! I personally love Go, but Rust's performance is truly impressive.

Glad you like it! Trying to make it as user-friendly as possible.

Yes! tree-sitter supports multiple languages and the core algorithms should transfer easily. I focused on Python first because I saw many people struggling with code quality issues in Python.

Not yet! But the algorithms should transfer well - the core logic supports TypeScript, so it's definitely doable.

Yeah Go is very fast!

I'm actually more optimistic. While Python/JS have huge ecosystems, there are still things only Go/Rust can achieve.