Live data from Hacker News

Claude Code now supports hooks

docs.anthropic.com

181–184 of 184 posts

Re: Claude Code now supports hooks

#181
post #114

Earlier quoted context omitted.

I think it's going to go the opposite way: we'll get a lot more custom made software, that fits exactly what a small customer needs. The code might be utter crap, the design might not be award winning, but it will be custom made to a degree that you can't customize your average Savas.

Yeah and then you’ll get hundreds of slightly different protocols formats and standards and nothing will talk to each other anymore without bespoke integration

LLMs can do that integration :)

Re: Claude Code now supports hooks

#182
one thing that is getting clear is that the gains from model enhancement is getting saturated

thats why we are starting to see a programming of ai, almost like programming building blocks

if there is a pathway for models to get smart enough to know when to trigger these hooks by themselves from system prompt or by default itself, then it wouldn't make sense to have these hooks

Re: Claude Code now supports hooks

#183

Earlier quoted context omitted.

> Stop using early returns in void functions! It’s in the clause.md you shouldn’t do that in this project! Reading CLAUDE.md (22 seconds, 2.6k tokens..) You’re absolutely right!

Honestly patterns like these that can have a tool fix the LLM quirks seems powerful.

No doubt, it’s just that I shouldn’t have to. Remind the tool of its own configuration file.

Re: Claude Code now supports hooks

#184

I've been playing with Claude Code the past few days. It is very energetic and maybe will help me get over the hump on some long-standing difficult problems, but it loses focus quickly. Despite explicit directions in CLAUDE.md to build with "make -j8" and run unit tests with "make -j8 check", I see it sometimes running make without -j or calling the test executable directly. I would like to limit it to doing certain…

Maybe a good way to think about this is that humans like usability/automations because even through we can handle more complicated manual operations, "abstracting away" certain operations helps us free up context. E.g. I like to be able to press F5 to reconfigure, build, run and attach debugger automatically from the IDE. If I had to manually to all this every time it would really clog up my context. So maybe try to make things easier for the AI.

Another example, it was messing up string dictionary keys all the time, even after CLAUDE.md etc. I realised, if it's annoying for the human to get dictionary string keys right (for humans it's error prone, humans prefer constants or propterties that will give instant feedback by the IDE if there is a typo). Ideally Claude would have an MCP which linting/indexing. E.g. if it edits a source file, without building/running it instantly receives errors by the editing tool. And Claude should ideally have some kind of code index in its context in the same way my indexer for Java/C++/TypeScript etc probably has a cached index of all symbols somewhere, allowing instant error detection. Maybe even running a fast language server MCP on each line edit may help claude.

Specifically regarding make -j8, I think an IDE MCP would help, where it should provide the same amount of ergonomy has a GUI IDE provides; e.g. quick shortcuts for common actions such as search, find-replace, build and even debug/set breakpoint, find references. Currently Claude is grepping, which works.

Is there an IDE-MCP somewhere; with language server support? That would be quite something!

Post reply on HN