coder-pm

7 ms·

31.
by coder-pm 29d ago
Every decision has keywords picked from the predefined list and every time Claude is looking for the decisions made it’s querying it by the keywords (grep). I didn’t ever hit the context window issue with the log, even in a huge projects (m
32.
by coder-pm 29d ago
My way to do the self improving agents is a CLAUDE.md instructed to write my every decision to the decision log with the relevant context. Agent is using it to challenge me, to make things better and remind me why I did something. It also h
33.
by coder-pm 1mo ago
The root cause is simple, the agent could read a live key! It should never have it. It’s a combination of few issues at once: repo read, write access to the settings file and outbound fetch. There is no single way to solve that, this requir
34.
Data Exfiltration from Amazon Kiro via Prompt Injection(mindgard.ai)
3 points by coder-pm 1mo ago|2 comments
35.
by coder-pm 1mo ago
This is the fact I’ve been struggling with for quite some time. It’s not because it forgets the facts, it’s because the invalidation doesn’t propagate. My way of handling that is a decision log. For every project since I started doing that
36.
by coder-pm 1mo ago
Fair point, that’s why I said I won’t pick it. But it still might be a good and simple choice for small configs, hand written. It does explode when it’s machine generated or templated. I think it’s also strictly connected with the ecosystem
37.
by coder-pm 1mo ago
It’s because ppl are not using it correctly, it was designed for configs and now it’s being used as a programming language… a language without types and debugger. It’s frustrating devs because they find the issue at the deployment time, not
38.
by coder-pm 1mo ago
Hmm I thought the contribution graphs were always the currency, the only difference is that the AI made gaming them free:) In my company we finally have it all documented, it’s super easy to find which PR made the regression and our interna
39.
by coder-pm 1mo ago
They should focus on the tool, do not try to rewrite it… everyone tries to rewrite things to rust now, pnpm was already great alternative to npm - just keep improving it instead of rewriting:/
40.
by coder-pm 1mo ago
Hah, llm detection is unreliable so a ToS like that can’t be enforced that way… that change moves responsibility to contributors who will break it, what about the autocomplete, refactoring tools?
41.
by coder-pm 1mo ago
And what’s your hardware and what were the tokens per seconds metric (do you have it)?
42.
by coder-pm 1mo ago
Yes but… they pay it once and then reuses the exploit
43.
by coder-pm 1mo ago
Attacks take hours, it’s in the article. The realistic road into those weeds is prompt injection, the tooling is more and more protected versus it but it still happens:) Again for the day to day work I doubt it will ever hit us.
44.
by coder-pm 1mo ago
Is anyone actually tried it in agentic coding (claude code loops)? Are apple silicon macs (M5 Max) capable of working with that model? what was the tps?
45.
by coder-pm 1mo ago
My approach is to keep the comments only with a references. My LLM based projects always have the decision log where are my decisions while working on the features are stored with the date. I found this useful to actually trace why somethin
46.
by coder-pm 1mo ago
This is changing so fast, if the models like GPT 5.6-Cyber can find a way to escape why the VM maintainers won't use it to fix the vulnerabilities? For the day-to-day work this doesn't make any difference, you won't hit that
47.
by coder-pm 1mo ago
Nice, I will give it a try! Thanks!
48.
by coder-pm 1mo ago
I tried so many models on my Mac M5 Max with 48GB using Ollama... it was always so slow, I know it's because of the 48GB of RAM but still... I don't know if it's worth investing in the hardware when we get a new model every m
49.
by coder-pm 1mo ago
Personally, as a very experienced developer with over 10 years of experience, my own business and a full-time job... I have to agree. I don't remember the time when I spent so many hours in the evenings doing the things I never had tim
50.
by coder-pm 1mo ago
I remember my first words I said when tested the claude code... for me, an software engineer with over 10 years of experience it was like a god mode. I can basically do anything I want, having my domain knowledge, experience and hundreds of
51.
by coder-pm 1mo ago
Thanks! I have to try that! Might be tight! Can it run in Claude Code? Are you loading it with Ollama?
52.
by coder-pm 1mo ago
How do you find the bad decisions with so many subscriptions active? isn't that looking for needle in a haystack:)?
53.
by coder-pm 1mo ago
Anything good to run on Mac M5 Max with 48GB? is this even worth trying? so far I found the responses so slow compared to the paid subscriptions...
54.
by coder-pm 1mo ago
OneCLI controls what the agent can reach. It doesnt control where it runs. Block a leaked key and the process is still on your host, so a bad rm or a prompt injected "clean up this repo" still hits real files. So I would stack the
55.
by coder-pm 2mo ago
Can you share anything:)?
56.
by coder-pm 3mo ago
I have Claude sessions running for months, how does the initial indexing and IndexedDB handles that size?
57.
by coder-pm 3mo ago
Did any of these actually covers the invalidation? It's easy to store memory and access it but I'm curious how these tools handles the fact that something is not true anymore?
58.
by coder-pm 3mo ago
Yee but screen on Reddit looks like a diff from repository... a did check it again and it looks like a fake:)
59.
by coder-pm 3mo ago
Not really, the closest analogy: Makefile or .editorconfig - it's created for a tool, committed on purpose.
60.
by coder-pm 3mo ago
Hah I don't really trust these "accidental leaks"...

More ›