Top
Best
New

Posted by MisterBiggs 2 days ago

You're probably using Agent Skills wrong(notes.ansonbiggs.com)
73 points | 25 commentspage 2
theowaway213456 2 days ago|
TL;DR don't have your agent write skills using only its latent knowledge, otherwise you may as well not use a skill in the first place and let it summon that latent knowledge on the fly.

Not sure if this take is correct though. I suspect self-generated skills help the agent avoid having to "decompress" its latent knowledge, which might save tokens? idk, I am not an expert

solarkraft 1 day ago||
It seems so obvious: How would it know better than it already does?

Yet I’ve seen people succeed with „write me a prompt“ prompts. The model makes something up, often it makes sense.

They are like plans in that way: It’s not exactly novel knowledge, but it at least encodes it somewhere to make the process verifiable beforehand and a bit more repeatable.

I wouldn’t be surprised if it improves performance a little, just like thinking blocks do (every model reasons now).

bigcat12345678 2 days ago|||
I now have rules to not let agent write any docs or processes. Pretty much anything LLM auto-generated are of zero reuse value.
imhoguy 1 day ago||
Autogenerated content is good scaffolding, but then I have a rule where if I mark heading with "(by-human)" the section shouldn't be changed by LLM without permission.
cassianoleal 1 day ago||
Skills can transfer one session's latent knowledge to all other sessions.
Atom_Foundry 1 day ago|
[dead]