- Tested by hand to refresh my brain and bash-fu and actually read the code
- tested on x86 Linux and macOS 27 dev beta ( oh boy…)
issue : 404 because v0.7.0 release does not exist
there is somewhere or somehow v0.7.0 and it fails ( at least it returns http status unlike some other 100k star repos )
I fixed it as any sane person would and I opened Github again, copy pasta, curl -OL <release-bin> ~/.local/bin (tested both aarch64 and x86
Then I got to read the script and practice my brain making comments ( yes, it's a bit silly.) and after that tried testing the actual replay
Happy to give feedback and propose some ideas and changes. I'm the FIRST FORK ! :)
I wanted to compile from src but got sidetracked by reading bash :D*
('ve noticed you mentioning limactl in the docs/ - I was just making a bunch of configs with all core eBPF tools to see if it might work in the latest Apple/container ( it does for the most part ) - to kill a rogue agent here and there or just have fun…
Happy to give it a shot and add that as a second exec backend, I prefer it over Docker by a long shot, because unlike Docker it actually uses real virtualization ( I think at least ? )
Some of our work overlaps here and there ( https://github.com/omry/arbiter and reploy ) + I got an engine that's a bit like a rewrite of Ansible to bun and Typescript with "agents-in-the-loop" , with 4 execution backends - docker, agent-sandbox ( CNCF/Google - pre-warmed pods on GKE ), k8s API, and sandlock ( truly fantastic and easy to use - it would fit perfectly into the suggested use cases of yours https://github.com/multikernel/sandlock ) .
—
Anyway , I dig the 3 projects you are working on , seems like a breath of fresh air ( the video !!! We'll need something like that for our company sooner or later , such a nice concept. ) . Found Hydra by randomly scouting Github on what's new ( thanks Zuck lol, came through opensource Meta web > Hydra haha )
Apologies for a long post, I'd just be happy to add more feedback + talk about SW lifecycle in this agentic era - human/agent(fleets) and how to make it work for people and good faith actors that can actually maybe contribute under certain conditions.
Cheers,
Hans
Disclaimer : no LLM used to write a single char in the text above. But otherwise I use it very heavily.
issue : 404 because v0.7.0 release does not exist
there is somewhere or somehow v0.7.0 and it fails ( at least it returns http status unlike some other 100k star repos )
I fixed it as any sane person would and I opened Github again, copy pasta,
curl -OL <release-bin> ~/.local/bin(tested bothaarch64andx86Then I got to read the script and practice my brain making comments ( yes, it's a bit silly.) and after that tried testing the actual
replayHappy to give feedback and propose some ideas and changes. I'm the FIRST FORK ! :)
Note:
I wanted to compile from src but got sidetracked by reading bash :D*
('ve noticed you mentioning
limactlin the docs/ - I was just making a bunch of configs with all coreeBPFtools to see if it might work in the latest Apple/container ( it does for the most part ) - to kill a rogue agent here and there or just have fun…Happy to give it a shot and add that as a second exec backend, I prefer it over Docker by a long shot, because unlike Docker it actually uses real virtualization ( I think at least ? )
Some of our work overlaps here and there ( https://github.com/omry/arbiter and
reploy) + I got an engine that's a bit like a rewrite of Ansible tobunand Typescript with "agents-in-the-loop" , with 4 execution backends - docker, agent-sandbox ( CNCF/Google - pre-warmed pods on GKE ), k8s API, andsandlock( truly fantastic and easy to use - it would fit perfectly into the suggested use cases of yours https://github.com/multikernel/sandlock ) .—
Anyway , I dig the 3 projects you are working on , seems like a breath of fresh air ( the video !!! We'll need something like that for our company sooner or later , such a nice concept. ) . Found Hydra by randomly scouting Github on what's new ( thanks Zuck lol, came through opensource Meta web > Hydra haha )
Apologies for a long post, I'd just be happy to add more feedback + talk about SW lifecycle in this agentic era - human/agent(fleets) and how to make it work for people and good faith actors that can actually maybe contribute under certain conditions.
Cheers,
Hans
Disclaimer : no LLM used to write a single char in the text above. But otherwise I use it very heavily.