Show HN: Mu – Tools for Agents | Hacker News Reader

Show HN: Mu – Tools for Agents

github.com

52 pointsasim26 comments

I think if nothing else the technical facing value proposition should be human written.

Communicating is hard but that’s the point, takes a lot of effort to close the gap between a new and skeptical user and your vision.

Devs will not be able to bypass the PTSD of reading Claude prose.

Amen. If i see a 5000 line README i am closing the tab.

I have in my CLAUDE/AGENTS.md to be concise with both chat output and documentation.

Yet still I get walls of text often. It's so fatiguing.

Root agents.md should be human authored. Even if you copy and paste into it. That's your only lasting thread into having any intuition on the behavior that will result

You're right about that and I totally hate it as well. I had originally written it myself and then it got revamped along the way as I tried to change the positioning of the system from a personal home server to tools for agents.

Good observation thank you.

but... why?

you almost never want these tools in any given session, and creating any of them can be done simply with Claude. looking at the PRs, it seems like that is exactly what's happening. i guess this is my generic problem with MCP releases, who are they for and why?

I’d go even further and add that these distinct tools don’t belong in the same toolbox.

Couldn't agree more.

We don't need large collections of assorted tools in one MCP server. Not trying to detract from the effort OP made... but we need a mechanism for finding, installing and managing our MCP servers effectively. Then monolithic collections aren't needed.

I believe collection servers are a band aide for the real problem. Useful for the individuals who built them, and often times not for too many others..... which is why I am building management tools for MCP servers. Hopefully to minimize the need for doing this. Because OP has filled a gap in the path available to them.

Im not OP but as someone who did one of these recently, I think about it kind of like sharing an essay. I had an idea, I clauded a working thing, used and refined until it was useful for my team then wanted to share the kernel that developed. OP is not claiming it’s like a panacea or a masterpiece and idk I appreciate the thoughts it’s making me think.

My other more general thought here is that the nature of LLMs is so generic and context specific that just learning about a workflow can unlock a ton of potential. This is evident in how the null state of every chat UI has a bunch of examples to teach you to use it. MCP’s are a portable way to share a workflow. To get concrete: I was pulling a lot of business data sources into my Claude code sessions to great effect but it was a force multiplier for my team to figure out an MCP version that would tell their Claude’s how to do the same. None of it was deeply technical engineering but it was immediately useful and most of the team using it are not engineers.

Thanks for that commentary. I think it's an evolution of an idea. So way back when I was working on micro services and apis. I had this idea of putting all the common API building blocks in one place. So rather than signing up to twilio and Auth0 and Google and whatever other API providers it would just be in one place and you would get this kind of streamlined simple way of accessing it. One token you could scope it to. I put that idea on the back burner when I ran out of money.

This project didn't start as tools for agents. It started as a personal app server where I could consolidate a lot of my daily habits that were apps. I actually wrote that by hand. Then I got access to Claude and it made life easier so I was able to extend it.

But I feel like my original goal was sort of unfulfilled and that's when I realised all of this stuff could actually be useful to agents and the same idea I had about apis for developers could apply to agents. So maybe it's not going to be something everyone wants to use. But for me I'm kind of interested to build this collection of let's say a hundred of the most useful tools put them behind one mCP server with access through one token and one account and then I'm going to start to kind of program against it. And the idea is well. If you give the agent access to real world things, you could build real world agents. And maybe people are already handcrafting this but then it's not made for that audience right. So I think that's why hacker news can be somewhat subjective. I've always found it quite polarising whenever I posted anything because you know there's this half that just say well. I don't need this because I can do it myself and then there's the other half they're like. Wow, this is cool.

Anyway, I appreciate the feedback

for high effort MCPs I see the reason. Big MCPs take major tokens to produce and test (if you're just telling an LLM to handle it) , it's nice if there is a high quality well tested thing out there to just plug into.

for things that are essentially a skill compendium, I see less reason..

contributors are one human and four agents, what an all star team

:D you know last time I built software go-micro was 4 years slogging it out on my own. Then raised funding and had 4-5 engineers working on it. Now you can honestly go very far with Claude, copilot, codex, etc. Still the person leading needs to understand what they're doing.

Here's the last one. https://github.com/micro/go-micro

I admire your grit Asim, you never stop :) This looks great !! but what are we using for searching news here, I mean the sources ?

Thank you! The news is an RSS reader/aggregator, so picking select sources to pull from. I opted not to use anything else as this works quite well and we don't need more news than we have. Plus we have Brave for web search. The sources are like the BBC, the guardian, techcrunch, hackernews, etc.

Hey, I liked this tool. Do you plan to develop more on the app creation tool? Is that shareable artifacts or what is the value prop?

Thanks. We can do. The idea was that it as more like WeChat or Telegram mini apps. So they are things you'd build in the system that have access to the system. I'd be curious to know your use case or what you want out of it.

Faith area is interesting.

Thank you for the interest and discussion all!

Funny how you made the name Micro / MU while there is a major company called Micron with the stock ticker MU. Intentional I guess?

I honestly don't really get MCPs. They're like all the bad design choices of LSP multiplied by 5

Not to be confused with Mü, the card game!

or Mu, the lost continent.

Nor the Justified Ancients of Mu-Mu.

Genuine question rather than a criticism: what do the tools actually execute in?

I spent last week asking people running local agents what their generated code runs inside, and the answer was consistently "nothing" - same filesystem, same network, same credentials as the agent itself. One person reviews everything with a second model and then reads it himself, which is the most careful answer I got and still isn't a boundary.

That's fine while a tool is reading a file. It stops being fine the moment the tool is "run this". Does Mu draw a line there, or is it left to the host?

By the way, it seems that your comments are getting automatically "killed" by the HN system. I don't see a clear problem with them, but I can see why people might think they've been AI generated, which is not acceptable here (this is somewhat buried in https://news.ycombinator.com/newsguidelines.html , which you should familiarize yourself with).

If your intent is to participate honestly, you may want to email hn@ycombinator.com to figure things out.

Reply on news.ycombinator.com