HackerLangs
トップ新着トレンドコメント過去質問紹介求人

mkagenius

3,296 カルマ登録 13 年前
https://instavm.io (manish [at] instavm.io)

投稿

Self-host sandbox cloud and hypervisor for AI agents

github.com
1 ポイント·投稿者 mkagenius·昨日·0 コメント

Show HN: Tarit – Self-host sandbox cloud and hypervisor for AI agents

github.com
6 ポイント·投稿者 mkagenius·3 日前·0 コメント

Burned out creating public software

mkagenius.substack.com
4 ポイント·投稿者 mkagenius·8 日前·1 コメント

Hey GLM 5.2, build me a hypervisor

technotes.substack.com
6 ポイント·投稿者 mkagenius·9 日前·0 コメント

How Claude Code and Codex approach sandboxing

instavm.io
2 ポイント·投稿者 mkagenius·11 日前·1 コメント

Taste and judgement are lies we tell ourselves

technotes.substack.com
2 ポイント·投稿者 mkagenius·15 日前·0 コメント

Taste and judgement are lies we tell ourselves

technotes.substack.com
5 ポイント·投稿者 mkagenius·21 日前·0 コメント

Show HN: Manim-Studio – Prompt to 3blue1brown video

github.com
3 ポイント·投稿者 mkagenius·26 日前·0 コメント

[untitled]

1 ポイント·投稿者 mkagenius·2 か月前·0 コメント

Show HN: See How You Talk to Claude Code

prophetic-tacky-bison.instavm.site
3 ポイント·投稿者 mkagenius·3 か月前·0 コメント

A retro terminal music player inspired by Winamp

github.com
145 ポイント·投稿者 mkagenius·4 か月前·47 コメント

Show HN: Cronbox – Schedule AI Agents

cronbox.sh
2 ポイント·投稿者 mkagenius·4 か月前·0 コメント

China's OpenClaw users paid to install viral agent. Now they spend to remove it

scmp.com
3 ポイント·投稿者 mkagenius·4 か月前·0 コメント

Hacking 'Skill' distilled from 4200 Hackerone vulns

instavm.io
2 ポイント·投稿者 mkagenius·4 か月前·0 コメント

Show HN: Run Claude Code in a Local Sandbox

github.com
1 ポイント·投稿者 mkagenius·4 か月前·0 コメント

A tape is all an agent needs

turing-s-tape-911416006313.us-west1.run.app
1 ポイント·投稿者 mkagenius·4 か月前·0 コメント

Show HN: Sandbox Flow – A Playground for Sandboxes

github.com
2 ポイント·投稿者 mkagenius·4 か月前·0 コメント

Music Maker: Automate Music Generation with Claude Code

github.com
1 ポイント·投稿者 mkagenius·4 か月前·0 コメント

I can no longer write technical blogs due to coding agents

gist.github.com
3 ポイント·投稿者 mkagenius·4 か月前·4 コメント

Show HN: Can this model run on my Mac mini?

mac-ai-model-compatibility-checker-911416006313.us-west1.run.app
2 ポイント·投稿者 mkagenius·4 か月前·0 コメント

コメント

mkagenius
·3 日前·議論
> Free idea: I would like to be able to "jail" an agent inside a VM and send instructions to the harness from outside the VM to agent(s) installed inside. Ideally there is no Codex/Claude/etc. installed on the host.

You can do exactly that with coderunner

https://GitHub.com/instavm/coderunner
mkagenius
·12 日前·議論
One would. But then the results are even weirder as opus 4.6 scored more than opus 4.8 by a huge margin
mkagenius
·12 日前·議論
It looks like the author is specifically avoiding model's name, because results are really weird.

  Opus 4.8/4.7 scored 28%

  Opus 4.6 score 37%

So the author thought as let's not get into that just write Claude.
mkagenius
·12 日前·議論
Token smuggler sounds like a profession coming soon. For distillation and stuff.
mkagenius
·14 日前·議論
Have you tried https://instavm.io in US
mkagenius
·14 日前·議論
Useful in splitting a big task - some parts are easy so give it to say Gemini. Some are harder so give it to gpt 5.5 and so on.

Also the throughput kind of increases since providers are different.
mkagenius
·14 日前·議論
We have created Murmur[1] which kind of works with your existing subscription (having API key is not mandatory). You can just tag @copilot @codex from claude code to delegate work to them. (it can also do it on its own too btw)

1. https://github.com/instavm/murmur - Murmur
mkagenius
·14 日前·議論
Not so subtle plug for another sandbox provider, https://instavm.io :

Apart from the above features.

  1. We support more than 32GB disk (as a shareable device, ideal for agentic memory)

  2. We provide egress control

  3. We provide vault for secret injection (to counter prompt injection)

  4. Snapshot / forking.

  5. long lived sandboxes.
Everything supported in APIs and CLI for agents.

Can be used via - npx skills add instavm/skills
mkagenius
·26 日前·議論
AWS for AI agents - https://instavm.io

Providing sandboxes through a CLI. Guardrails such as egress control and secret injection and audit trails built in.

We can also be used as 3rd party sandboxes in Anthropic managed agent and OpenAI sdk.

https://instavm.io/blog/self-hosting-claude-managed-agents-o...
mkagenius
·27 日前·議論
6 months a bummer, but we got it for apple sandbox - coderunner (https://github.com/instavm/coderunner)

We got it yesterday, maybe they just started rolling it out and hence op posted this.
mkagenius
·28 日前·議論
Yes it is funnily true but it was for fake news generation and not it's cyber capabilities.

Another fun little gem of information, government has something called Mayhem

> the autonomous Robo-Hacker AI called Mayhem that’s now in charge of protecting the Pentagon’s most critical systems

Guess Mythos and Mayhem had a chat
mkagenius
·28 日前·議論
[flagged]
mkagenius
·先月·議論
Apple containers are great for providing a sandbox to your AI coding agents

I have made it a MCP so that it's easily discoverable by all the coding agents

https://github.com/instavm/coderunner
mkagenius
·先月·議論
[dead]
mkagenius
·2 か月前·議論
The article misses the point that you can have a third option - harness inside and outside both.

The use cases will evolve so much more in coming weeks and months that the boundaries will blur. You can still contain the agent outside in a sandbox though.
mkagenius
·3 か月前·議論
With instavm (https://instavm.io), you can provide an OCI image built from a dockerfile.
mkagenius
·3 か月前·議論
I also felt the need for a cloud based cron like automations, so decided to build it myself https://cronbox.sh with:

  1. an ephemeral linux sandbox for each task 
  2. capability to fetch any url
  3. can use tools like ffmpeg to fulfill your scheduled task
mkagenius
·3 か月前·議論
Not just sides, the vents are much sharper.

Few years back, I tried to look on reddit for complaints regarding this - barely anything.
mkagenius
·3 か月前·議論
> The biggest problem is the coding agents don't "Fail fast and loud". They fail deceivingly.

GPT 2 and 3 used to fail fast (and loud coz we could easily see it lying)
mkagenius
·3 か月前·議論
Has to be taken in light of median income too right?