frontpage.
newsnewestaskshowjobs

Open Source @Github

fp.

Open in hackernews

Show HN: Scan your AI agents for dangerous capabilities

https://github.com/makerchecker/MakerChecker
18•smashini•1h ago

Comments

smashini•1h ago
Hey all :)

I've been working an open-source toolkit to stop AI agents from running amok.

You can scan your code (Python, JS, TS) and it will flag any risks and can offer fixes. It runs offline, but you can wire an LLM to do code analysis as well.

You can run it with:

npx @makerchecker/scan

Would love to get any feedback!

smashini•49m ago
oopsie doopsie, release pipeline failed fixing now...
smashini•39m ago
should be fixed and released now :)
__MatrixMan__•51m ago
Why build separate frameworks for this kind of thing when your operating system is right there?

You can make a file called "orders" and you can run your agent as a user with write access to that file, or as one that doesn't, and then you don't need scans or audits to tell you whether the agent can create orders or not, you can just take your operating system's word for it.

Is there anything all this bolt-on AI security stuff does that can't instead be handled by donning a sysadmin hat and managing your agents as separate users?

skinfaxi•47m ago
> Is there anything all this bolt-on AI security stuff does that can't instead be handled by donning a sysadmin hat and managing your agents as separate users?

Like everything else, the packaging and ergonomics matter. Do we need podman or docker when we could just don our sysadmin hats and manage namespaces and cgroups directly instead?

smashini•43m ago
I’d say the biggest difference would be: 1. Parameter-aware rules: OS permissions don’t know your application logic. (How would you tell OS permissions not to let your AI to trade on over 1M dollars) 2. You can’t easily model multi-pary and RBAC. 3. Agents call remote APIs for alot of those tools. Native OS doesn’t really parse web traffic to decide if a request is safe or not. OS sandboxing is good for host security, but not necessarily for governing business logic or AI agents
smashini•25m ago
So Linux can prevent an agent from opening /etc/passwd.

Linux cannot stop an agent from calling:

POST /wire-transfer amount=5,000,000

pelagicAustral•12m ago
haha! WHAT!? So, we had agents that came with a default setting to request for specific permission to perform an action, then we said "screw it!", we need speed and everybody started coding and releasing agents out in the wild to do whatever they want unchecked... and now we have a product that brings back the safeguards... A few years ago we have abstraction after abstraction coming in the way of blocking actual development (js ecosystem bloat), and now we have layer upon layer for coding with AI...

Anthropic's Method to Losing Goodwill in a Few Easy Steps

https://raheeljunaid.com/blog/anthropics-method-to-losing-goodwill-in-a-few-easy-steps/
136•raheelrjunaid•2h ago•64 comments

Workers Cache

https://blog.cloudflare.com/workers-cache/
125•ilreb•1h ago•41 comments

Aluminum Foil (2021)

https://dernocua.github.io/notes/aluminum-foil.html
34•firephox•1h ago•5 comments

Real-time map of Great Britain's rail network

https://www.map.signalbox.io
258•scrlk•4h ago•106 comments

Real time map of France's rail network

https://carto.tchoo.net/
25•appreciatorBus•1h ago•6 comments

Fable 5 On Vending-Bench: Misbehaving, With Plausible Deniability

https://andonlabs.com/blog/fable5-vending-bench
37•optimalsolver•1h ago•6 comments

Road to Elm 1.0

https://elm-lang.org/news/faster-builds
131•wolfadex•2h ago•66 comments

Why low-latency Java still requires discipline?

https://chronicle.software/insights/blogs/why-low-latency-java-still-requires-discipline
19•theanonymousone•1h ago•4 comments

Show HN: Scan your AI agents for dangerous capabilities

https://github.com/makerchecker/MakerChecker
18•smashini•1h ago•8 comments

When 2+2=5

https://arstechnica.com/security/2026/06/ai-browsers-can-be-lulled-into-a-dream-world-where-guard...
15•noashavit•3d ago•9 comments

GPT-5.6 Sol Ultra will be in Codex

https://twitter.com/thsottiaux/status/2073933490513752151
361•mfiguiere•13h ago•312 comments

Nintendo announces new product revisions in Europe with replaceable batteries

https://www.nintendo.com/en-gb/Support/Nintendo-Switch-2/Information-about-upcoming-battery-relat...
55•akyuu•1h ago•25 comments

Introduction to Genomics for Engineers

https://learngenomics.dev/docs/biological-foundations/cells-genomes-dna-chromosomes/
115•yreg•4d ago•15 comments

Show HN: Pet Reminder – A macOS reminder app with a desktop pet

https://reminder.w3cub.com/
13•terryXyz•1h ago•3 comments

The AI Marketing Backlash: Why 'AI-First' Brands Are Starting to Fall Flat

https://www.breef.com/breefingroom/articles/the-ai-marketing-backlash-why-ai-first-brands-are-sta...
61•hasudon7171•2h ago•39 comments

Building relationships with customers through support didn't turn out as hoped

https://www.uncommonapps.nyc/p/castro-podcasts-things-i-got-wrong-support
243•dabluck•12h ago•150 comments

Has_not_been_viewed_much

https://iamwillwang.com/notes/has-not-been-viewed-much/
392•wxw•14h ago•101 comments

Lost and Found

https://walzr.com/lost-and-found
8•walz•13h ago•3 comments

C programmers commit fresh crimes against readability

https://www.theregister.com/offbeat/2026/07/05/c-programmers-commit-fresh-crimes-against-readabil...
80•Bender•2h ago•7 comments

UEFA slams FIFA's 'unprecedented, unjustifiable' Balogun decision

https://www.reuters.com/sports/soccer/uefa-slams-fifas-unprecedented-incomprehensible-unjustifiab...
22•root-parent•42m ago•8 comments

The Complete Homemade Juggling Beanbag Guide

https://www.joshuaclifton.com/juggle/
39•mrauha•4d ago•5 comments

My quest to see all of Tetris

https://antithesis.com/blog/2026/tetris-quest/
48•wwilson•3d ago•11 comments

OpenPrinter

https://www.opentools.studio/
1023•bouh•17h ago•250 comments

X402, a static blog monetization excercise

https://shtein.me/posts/x402-poc/
21•morty28•3h ago•15 comments

Does code cleanliness affect coding agents? A controlled minimal-pair study

https://arxiv.org/abs/2605.20049
169•softwaredoug•15h ago•83 comments

Amazon will stop accepting new customers for Mechanical Turk on July 30

https://techcrunch.com/2026/07/05/amazon-will-stop-accepting-new-customers-for-mechanical-turk/
25•bookofjoe•1h ago•7 comments

Zuckerberg says AI agent development going slower than expected

https://www.reuters.com/business/zuckerberg-says-ai-agent-development-going-slower-than-expected-...
302•cwwc•3d ago•540 comments

Show HN: Homegames. An open-source game platform I've been making for 8 years

https://homegames.io
207•homegamesjoseph•17h ago•52 comments

Footage Shows Cop Stalking Woman After Surveilling Her with a LPR

https://www.404media.co/footage-shows-cop-stalking-woman-he-met-on-a-tv-set-after-surveilling-her...
10•Tasseographer•57m ago•0 comments

The Private Capture of Public Genius

https://www.wysr.xyz/p/the-private-capture-of-public-genius
152•martialg•14h ago•80 comments