https://nf-co.re logo
Join Slack
Powered by
Implementation PR for the agents RFC is here: <htt...
# team-maintainers
i
Implementation PR for the agents RFC is here: https://github.com/nf-core/proposals/pull/143 Feel free to have a look and share your thoughts. I will start working on the drafts soon, but I want to get as much feedback as possible before we get deep into the implementation.
👀 1
🎉 1
@Evangelos Karatzas, @Phil Ewels: feel free to comment and/or participate in drafting <!subteam^S016E36N7TQ>: point about new repository for your consideration <!subteam^S08LAB76NBU>: parts of this RFC are relevant to you
p
Thanks @Igor Trujnara! Left a review.
i
Thanks, replied now
j
Very nice @Igor Trujnara! Left some very small suggestions, nothing very important
i
Everything fixed @Jose Espinosa-Carrasco
❤️ 1
Seeing that there are no more comments, are we okay to proceed with creating the new repo so I can start drafting the main 
AGENTS.md
? <!subteam^S016E36N7TQ>
✅ 2
m
Please wait with creating the repo. We need to add it to the ignore list first
p
Are we definitely happy with the name 
nf-core/agents
?
m
I would prefer llm-tooling or so. Because it will probably also contain skills etc
👍 1
Also making it a more Future proof
m
llmtools?
this 1
j
nf-claw
😁 1
shake no 1
🦞 1
p
nf-core/ai-tooling
?
👎🏻 1
👎 1
m
nf-core/contribute
m
nf-core/spellbook😉
🧙 4
😆 1
🧙🏻 1
nf-core/MI6 🕴️
m
nf-core/terminator
p
ok nf-core/agents is fine then 😆
😁 1
m
nf-core/llm-files?
j
nf-core/terminator
I HAD THR SAME IDEA
❤️ 1
m
(it should be nf-core/skynet then 😉)
party this 1
p
llm-files is fine but just a bit flat. Could call other repos things like "pipeline-files-for-rnaseq-analysis" by that logic 😅 Nicer to follow the purpose (eg. agents)
j
I like `llm-tools`tge most so far
m
(it's not just agents, plus agents is the old-school term. workers is the newer shit...
> I like llm-toolstge most so far too close to 
nf-core/tools
that's why i would prefer 
nf-core/llm-tooling
👍 1
m
nf-core/tools should be nf-core/cli to be honest
100 rainbow 1
j
llm-toolkit
m
It's not really a tool kit
i
tooling
does not suggest LLMs/agents to me at all
m
sorry meant 
llm-tooling
👍 1
m
For me it's really more guidelines about how to contribute, for anyone, agents included, so nf-core/contribute
But not #C0749PAEU9F
m
nf-core/tools should be nf-core/cli to be honest
it has the template files as well (which should be in their own repo, but it's nice to have some historical ballast)
nope, that content is not really for human consideration. It's just that LLMs don't follow coding standards and just check standard files by default.
👍 1
👍🏻 1
this 1
i
I support the no-hyphen ideology, but also not #C0749PAEU9F
p
nf-core/llms ?
👎🏻 1
m
we don't provide llms there...
nf-core/co-co-scientist to follow the newer buzz winking cowboy
🫠 2
j
nf-core/assistents
or 
nf-core/context
👆🏻 1
i
I saw the term "steering files", but I'm not sure if it's adopted outside Amazon
j
I think that even if it has an hyphen something such as agents-assistents or agents-context is more informative
this 1
j
nf-core/assistents or nf-core/context
Too generic, we need to have ai or llms in there
this 1
m
nf-core/botler? Y’know, mixing bot and butler 😅
😆 2
i
nf-core/aiagents
nf-core/aitooling
nf-core/aisteering
nf-core/aicontext
One of these could reconcile all parties, what do you think?
j
I like aicontext Steering is bad because we have the steering committee
this 1
m
ai is a marketing term and too broad
ai thumbsdown 1
j
llmcontext then
m
nf-core/aiaiaiaiai, and the logo is a mariachi band
🇲🇽 4
100 rainbow 2
😆 5
p
ai is a marketing term and too broad
I disagree, I think it's the main term used by the majority of people to describe LLMs
💯 2
but honestly I'm reaching exhaustion level on this thread so don't really care much any more
thistbh 1
kind of regret mentioning it as we probably would have just gone with nf-core/agents if I didn't say anything 😅
😂 3
j
For what it's worth I think 
nf-core/agents
is concise and descriptive in that you know what to look for
👍🏻 2
agree 3
👍 3
Open in Slack
PreviousNext
Powered by