Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

Claude Skills are awesome, maybe a bigger deal than MCP

simonwillison.net

Post reply on HN
Print view View original

301–310 of 383 posts

Re: Claude Skills are awesome, maybe a bigger deal than MCP

#301

Post by onel »

michael1999 wrote: post #203

We're doing something like this internally. Our monorepo context files were much too big, so we built a progressive tree of fragments to load up for different tasks. I am struck by how much these kinds of context documents resemble normal developer documentation, but actually useful and task-oriented. What was the barrier to creating these documents before? Three theories on why this is so different: 1) The feedback…

I've built an agent that builds documentation for code bases and you are 100% right. Having big picture documentation is important, but having bite size explanations about why some components work a certain way is more important. Making sure the AI doesn't have to infer behavior from code is really powerful. Even going as low level of reference docs. Even though devs would prefer that a method be self-Explanatory, it helps to also have plain english explanation about what's happening in a class or method.

Re: Claude Skills are awesome, maybe a bigger deal than MCP

#302

Post by AmbroseBierce »

They are awesome when you made them and work in the current version of things, I'm sure they can quickly become obsolete and the skills could soon no longer apply and you have to sit down and check what skill is failing and why, so it becomes another layer more of technical debt.

Re: Claude Skills are awesome, maybe a bigger deal than MCP

#303

Post by thenanyu »

jychang wrote: post #287
evanmoran wrote: post #280

Earlier quoted context omitted.

I know this is highly controversial, but I now leave the comments in. My theory is that the “probability space” the LLM is writing code in can’t help but write them, so if i leave them next LLM that reads the code will start in the same space. Maybe it’s too much, but currently I just want the code to be right and I’ve let go of the exact wording of comments/variables/types to move faster.

I think the code comments straight up just help understanding, whether human or AI. There's a piece of common knowledge that NBA basketball players can all hit over 90% on free throws, if they shot underhand (granny style). But for pride reasons, they don't throw underhand. Shaq just shot 52%, even though it'd be free points if he could easily shoot better. I suspect there's similar things in software engineering. I'…

I hate this analogy. NBA players can all hit 90% of their free throws shooting overhand too. Just some of them are much worse at handling the pressure and pace change of the situation in a game context.

Re: Claude Skills are awesome, maybe a bigger deal than MCP

#305

Post by AmbroseBierce »

rane wrote: post #289

So far not impressed with CC's ability to invoke skills automatically. I made a skill with the unambiguous description: "Use when creating or editing bash scripts" Yet, Claude does not invoke the skill when asked to write a bash script. https://gist.github.com/raine/528f97375e125cf97a8f8b415bfd80...

Try writing it like this: "Use it when creating a bash script or editing a bash script"

Re: Claude Skills are awesome, maybe a bigger deal than MCP

#306

Post by eloisant »

svdr wrote: post #167
phs318u wrote: post #149

Earlier quoted context omitted.

Not sure why you’re getting downvoted. The marketing that LLM=AI seems to have been interpreted as “_only_ LLM=AI”

I think the difference now is that traditional software ultimately comes down to a long series of if/then statements (also the old AI's like Wolfram), whereas the new AI (mainly LLM's) have a fundamentally different approach.

You're talking about non-deterministic algorithms, who yes are often associated with AI but existed way before LLM's

Re: Claude Skills are awesome, maybe a bigger deal than MCP

#307

Post by rco8786 »

simonw wrote: post #37
rco8786 wrote: post #20

Earlier quoted context omitted.

I don’t really see how this replaces MCP tbh. MCP gives the LLM access you your APIs. These skills are just text files with context about how to perform specific tasks.

You don't need MCP if you can instead drop in a skill markdown file that says "to access the GitHub API, use curl against api.github.com and send the GITHUB_API_KEY environment variable in the authorization header. Here are some examples. Consult github-api.md for more."

I would much, much rather provide discreet APIs directly to the LLM via MCP than just tell it to hit the api and figure it out from the docs.

Re: Claude Skills are awesome, maybe a bigger deal than MCP

#308

Post by scotty79 »

michael1999 wrote: post #203

We're doing something like this internally. Our monorepo context files were much too big, so we built a progressive tree of fragments to load up for different tasks. I am struck by how much these kinds of context documents resemble normal developer documentation, but actually useful and task-oriented. What was the barrier to creating these documents before? Three theories on why this is so different: 1) The feedback…

Program will evolve and documentation will go out of date before it's gonna be read a single time. So a variant of 1)

Personally I don't write documentation because I don't read documentation. It's too often pure garbage, less reliable than 2023 LLM output so I just skip it and go to the source code.

I would read documentation written for AI because I know for a fact that it describes the thing accurately enough if the system works. Human addressed documentation almost always has no verification.

Re: Claude Skills are awesome, maybe a bigger deal than MCP

#309

Post by simonw »

hansmayer wrote: post #285
simonw wrote: post #158

Earlier quoted context omitted.

> How the hell can you predict they will "take off even more" when the feature is accessible for barely 24 hours at this point? That's what a prediction IS. If I waited until the feature had proven itself it wouldn't be much of a prediction. The feature has also been live for more than 24 hours. I reverse-engineered it a week ago: https://simonwillison.net/2025/Oct/10/claude-skills/ - and it's been invisibly powering…

> That's what a prediction IS. If I waited until the feature had proven itself it wouldn't be much of a prediction. No, that's merely guessing mate. Predictions are, at least in modern meaning, based on at least some data and some extrapolation model that more or less reliably predicts the development of your known dataset into future (uknown) values. I don't see you presenting either in your post, so that's not pred…

I think you and I are operating from different dictionaries. What you're describing is more what I'd call a hypothesis or maybe a forecast. I'm comfortable with my use of "prediction" to mean the same thing as a guess.

Re: Claude Skills are awesome, maybe a bigger deal than MCP

#310

Post by simonw »

rane wrote: post #289

So far not impressed with CC's ability to invoke skills automatically. I made a skill with the unambiguous description: "Use when creating or editing bash scripts" Yet, Claude does not invoke the skill when asked to write a bash script. https://gist.github.com/raine/528f97375e125cf97a8f8b415bfd80...

Hah, yeah that's a total miss there.

Maybe it messed that up because writing bash scripts is so core to how Claude Code works? Much of the existing system prompt (and I bet a lot of the fine-tuning data) is about how to use the Bash tool.

Post reply on HN