JyB

No one think any form of "prompt engineering" "guardrails" are serious security measures right?

Assuming you're using GitHub or similar, make pushing branches and creating PRs part of their prompts, and review on the GitHub app or equivalent? Seem like an orthogonal problem to those LLMs tools.

> I can start a Claude Code session and if I need to leave my laptop, I can respond from my phone anywhere

I've been looking for this for some time now. This is amazing if it delivers.

That's a funny response I have to admit. No 100% serious, I don't see the advantage.

That's exactly how you should do it. You can also plug in an MCP for your CI or mention cli.github.com in your prompt to also make it iterate on CI failures.

Next you use claude code instead and you make several work on their own clone on their own workspace and branches in the background; So you can still iterate yourself on some other topic on your personal clone.

Then you check out its tab from time to time and optionally checkout its branch if you'd rather do some updates yourself. It's so ingrained in my day-to-day flow now it's been super impressive.

I’m really wondering what internally causes this. No one likes it when they have outages, but it keeps happening. Is this a culture thing? Like pressure to ship features fast? Some under staffed teams or lack of ownership on some crucial components?

Because iterating multiple sessions through multiple terminals is obviously more efficient and seamless than interacting thought a scuffed IDE side panel ui.

Better prompts?

Same question. I don’t understand what they mean by that. It obviously seem pretty central to how Claude Code is so effective.

How can it not be exactly what’s happening?

I thought Claude code was pay as you go!?

There is something much deeper going on when you force yourself to actually write things down. This is especially relevant in engineering. That is why "RFCs" are so prevalent in many tech companies. They are often just as useful to the writer as they are to the reviewers.

There is the Claude Code cli, now Gemini CLI. Where is ChatGPT CLI?

> "LLMs are one of the technologies that exist, like autocorrection and CAD"

Pretending it's on par with these. No need for more. Seem inevitable enough.

Incredible that so many people still don't understand what an LLM is. Especially ones that you would expect to grasp it.

Yes apparently which makes this report pretty flimsy.

Therefore what are the odds that this is just the LLM doing its thing versus "a vulnerability". Seem like a pretty obvious bet.

I don’t see anything here that would prevent a LLM from generating these. Right?

I believe it is extremely important to disclose that the ‘responses leaks’ you obtained did not originate from LLM models themselves, but rather through other insecure systems / in a more conventional manner.

Just to avoid yet another case of hallucinations outputs getting misinterpreted.

It doesn’t have to be as performant nor fast. It can work and iterate alone when setup properly. All time spent is purely bonus. It is already inevitable.

Fair. I'm using Claude Code which is pay as you go. The Market will probably do its things. (The company pays anyway obviously)

Ive never used that many. The LLM performances collapse/degrade significantly because of too much initial context? It seems like MCP implems updates could easily solve that. Like only injecting relevant servers for the given task based on initial user prompt.

That autocompounding aspect of constantly refining initial prompts with more and more knowledge is so interesting. Gut feeling says it’s something that will be “standardized” in some way, exactly like what MCP did.

I guess I don't see the technical limitation. Seem like a protocol update issue.

Do you mean that some tools (MCP clients) pass all functions of all configured MCP servers in the initial prompt?

If that's the case: I understand the knee-jerk reaction but if it works? Also what theoretically prevents altering the prompt chaining logic in these tools to only expose a condensed list of MCP servers, not their whole capabilities, and only inject details based on LLM outputs? Doesn't seem like an insurmountable problem.

> It demands too much context.

This is solved trivially by having default initial prompts. All majors tools like Claude Code or Gemini CLI have ways to set them up.

> You pass all your tools to an LLM and ask it to filter it down based on the task at hand. So far, there hasn't been much better approaches proposed.

Why is a "better" approach needed? If modern LLMs can properly figure it out? It's not like LLMs don't keep getting better with larger and larger context length. I never had a problem with an LLM struggling to use the appropriate MCP function on it's own.

> But you run into three problems: cost, speed, and general reliability

- cost: They keep getting cheaper and cheaper. It's ridiculously inexpensive for what those tools provide.

- speed: That seem extremely short sighted. No one is sitting idle looking at Claude Code in their terminal. And you can have more than one working on unrelated topics. That defeats the purpose. No matter how long it takes the time spent is purely bonus. You don't have to spend time in the loop when asking well defined tasks.

- reliability: Seem very prompt correlated ATM. I guess some people don't know what to ask which is the main issue.

Having LLMS being able to complete tedious tasks involving so many external tools at once is simply amazing thanks to MCP. Anecdotal but just today it did a task flawlessly involving: Notion pages, Linear Ticket, git, GitHub PR, GitHub CI logs. Being in the loop was just submitting one review on the PR. All the while I was busy doing something else. And for what, ~1$?

It’s the whole “MCP” moto right? Basically keep re-feeding in each subsequent prompts the additional requested context?

Or US/Norwegian forces covered by NATO exercises.

The whole point of Datadog is that you can seamlessly go between products for the same event source. Doesn’t it just highlight a setup issue?

This is such a great move. Faith restored in the language after the generics debacle.