# Rules and requirements for LLM/AI assisted projects _Showcase · started by Zenith on Thu, Jul 30, 2026 6:57 PM_ --- ## Original post **Zenith** · Thu, Jul 30, 2026 6:57 PM Due to the influx of LLM assisted shovelware, these are the rules that are now required if you use LLMs either in part or full for development to promote your project here. - You must demonstrate that you have a clear knowledge of your project's codebase. This one is very simple. The most immediate sign that a project is LLM shovelware is if you can't even bother to write your own README. If I see — a slew of — em dashes everywhere, "It's not ABC — it's X, Y, and Z with DEF", 🚀 emoji and table vomit that LLMs love, and the usual slop, I can more or less gleam in 30 seconds your project is shovelware. This is embarrassing. There are even ways to deslop LLM output if you can be assed to put in the bare minimum effort. https://github.com/realrossmanngroup/no_ai_slop_writing_rules Another one is clear functional design patterns. A common design pattern used by LLMs that makes things borderline malicious is excessive announces. The number of projects spewed out by LLMs that love to try announcing every 10 seconds and for some bizzare reason shoving *everything in the app_data* of an announce I can't even count. If your project is a LLM assisted port of the Reticulum Python implementation, it must be compatible with the Python reference implementation. If it can't even work with one of the scripts from the Examples folder then go wrangle your LLM until it does. There are like 6 LLM tooled Rust "ports" sitting on Github that can barely handle a link, despite reporting "feature complete". The next one: - Your project must have 6 months of commit history Most LLM projects I see particularly in relation to Reticulum are abandoned after the first few months with little support afterwards. 6 months of commit history is more than enough to demonstrate that you are committed to maintaining your project, even if that is through the use of LLMs exclusively. And lastly: - Your project or post should have an LLM disclaimer usage. That's it. --- ## Reply 1 **bergie** · Thu, Jul 30, 2026 7:22 PM As the author ("editor"? "curator"?) of an LLM-assisted Reticulum port, I find these rules fair. Of course a bummer that this means reticulum-js won't be listed here in next five months, and instead the broken-and-unmaintained rns.js is. But at the same time, the five months will give me time to get much further with the port. Now we're pretty good with running a "leaf node", but I do want to get transport node capabilities in. And that also means looking at all those new routing improvements that landed in Python rns 1.4. And in that time I should have a few more reticulum-js powered projects up and running on top of the existing Signal K and Yjs integrations. In that timescale we will also be sheltering from the southern Pacific cyclone season, which means I will have at least a few weeks of dedicated time to work on this instead of sailing around. --- ## Reply 2 **Zenith** · Thu, Jul 30, 2026 7:40 PM **bergie** wrote: > As the author ("editor"? "curator"?) of an LLM-assisted Reticulum port, I find these rules fair. > > Of course a bummer that this means reticulum-js won't be listed here in next five months, and instead the broken-and-unmaintained rns.js is. > > But at the same time, the five months will give me time to get much further with the port. Now we're pretty good with running a "leaf node", but I do want to get transport node capabilities in. And that also means looking at all those new routing improvements that landed in Python rns 1.4. And in that time I should have a few more reticulum-js powered projects up and running on top of the existing Signal K and Yjs integrations. > > In that timescale we will also be sheltering from the southern Pacific cyclone season, which means I will have at least a few weeks of dedicated time to work on this instead of sailing around. Yes its fine you can post updates about reticulum-js no worries --- ## Reply 3 **Loelin** · Thu, Jul 30, 2026 8:57 PM This is why global ret blackhole management is a must against the ever increasing shovelware across transport nodes. Hyperscaling agents do not respect the Zen manifesto (as bandwidth is finite). ---