frontpage.
newsnewestaskshowjobs

Made with ♥ by @iamnishanth

Open Source @Github

fp.

Open in hackernews

Show HN: Agent-browser-shield – free extension to protect AI agents on the web

https://github.com/pixiebrix/agent-browser-shield
5•tschiller•1h ago
I've been experimenting with Claude Code, ChatGPT Agent, and OpenClaw to perform more open-ended tasks for me online. A big blocker I've hit on shopping and research tasks is the agent getting a key piece of info wrong.

For example, in one case, my agent decided to add a brand I don't like to the cart because the site flagged it as almost sold out

The HN crowd is probably pretty aware of the threats and can avoid them while browsing. But what about their agents?

I tried prompting, but it was ineffective, because once the AI saw the threat, it polluted/distracted its context

Looking at the research, I came across a couple of papers, SusBench and Decepticon. The Deception research benchmarks indicate that increased reasoning can perform worse, because the model rationalizes the dark pattern

So it seems the best approach has to be removing the information before it can pollute/poison the context

In my day job, we have a browser extension that started as a productivity extension. However, contact centers started using us for neutralizing insider, fraud, and social engineering threats.

So my team set out to create a browser extension to neutralize all the threats AI agents face

We're focusing on open-ended tasks, because the best practice for routine tasks is to have the agent script repeat actions

It's also a tricky area since AI agents view the web in different ways: DOM, a11y tree, and visually. So we needed to account for those differences in how we detect and neutralize threats

The extension we created is agent-browser-shield, which defends against three primary threats:

- Prompt Injection - Dark Patterns - Context Pollution

It's free and source-available on GitHub, ClawHub, and the Chrome Web Store: https://github.com/pixiebrix/agent-browser-shield

We plan on making an enterprise version that pairs with our low-code engine for letting teams easily create custom rules for business-specific sites and internal tools

Looking forward to feedback! Especially curious if anyone has agent traces that got poisoned or sites to red team against!

Comments

britt_joienr•56m ago
This will help save on tokens too since it's stripping out junk!

Facing life-threatening miscarriage in Arkansas, calls to governor didn't help

https://www.cnn.com/2026/05/28/health/arkansas-miscarriage-abortion-ban-propublica
1•orwin•47s ago•1 comments

The Relaunch of the Old West and Why I Chose Vanilla PHP

https://blog.alexseifert.com/2026/06/02/the-relaunch-of-the-old-west-and-why-i-chose-vanilla-php/
1•SeenNotHeard•1m ago•0 comments

Batching API Calls

https://www.mostlypython.com/batching-api-calls/
1•speckx•2m ago•0 comments

Show HN: Mashines.dev – Live-migrate microVMs between hosts without restarting

https://mashines.dev/
1•ktaraszk•4m ago•0 comments

Impermeabiliza uses AI to modernize waterproofing in Valencia

https://impermeabiliza.net/
1•ignival•7m ago•0 comments

Microsoft unveils new AI models

https://microsoft.ai/models/
4•helloplanets•8m ago•0 comments

Tesla Cybertruck resale value plunges amid sales slump

https://www.msn.com/en-us/money/companies/tesla-cybertruck-resale-value-plunges-amid-sales-slump/...
4•petethomas•9m ago•1 comments

AI enthusiasts racing against time; AI skeptics are racing against entropy

https://charitydotwtf.substack.com/p/ai-enthusiasts-are-in-a-race-against
1•SLHamlet•10m ago•0 comments

AgentSight: System-wide AI agent tracing and monitoring with eBPF

https://github.com/eunomia-bpf/agentsight
2•tanelpoder•11m ago•0 comments

I believe a whole generation of developers miss how open source used to work

https://twitter.com/mitsuhiko/status/2062181646804152626
3•tosh•11m ago•1 comments

Gooey: A GPU-accelerated UI framework for Zig

https://github.com/duanebester/gooey
3•ksec•12m ago•0 comments

Network State Propaganda

https://capirates.substack.com/p/theyre-telling-you-exactly-what-theyre
3•tjcrowley•12m ago•0 comments

The 15-minute city is a dead end

https://blogs.lse.ac.uk/covid19/2021/05/28/the-15-minute-city-is-a-dead-end-cities-must-be-places...
1•Anon84•14m ago•0 comments

Book of Cron Job [fiction]

https://www.nature.com/articles/d41586-026-01716-0
2•apotheosized•14m ago•0 comments

Show HN: Hive Trust – Ed25519-signed benchmarks for every AI inference primitive

https://thehiveryiq.com/trust/
1•thehivery•16m ago•0 comments

Knowable – Open-Source Personal AI Tutor on macOS

https://knowable.ca/
2•samuelzxu•17m ago•1 comments

See SBA Loans Around You

https://www.loanround.com
1•zarie•17m ago•0 comments

Safe Made Easy Pt.2: Don't Fear the Ref

https://ergeysay.github.io/safe-made-easy-pt2.html
2•ergeysay•18m ago•0 comments

I benchmarked Opus 4.8 vs. GPT 5.5 on 2 open source repos

https://www.stet.sh/blog/opus-48-vs-gpt-55-vs-opus-47-vs-composer-25
3•bisonbear•18m ago•0 comments

The Download: Trump's new AI order, and smart glasses for warfare

https://www.technologyreview.com/2026/06/03/1138322/the-download-trump-ai-order-smart-glasses-war...
1•joozio•19m ago•0 comments

10M requests in my bot black hole

https://gladeart.com/blog/10-million-requests-in-my-bot-black-hole-here-is-some-information
1•itsJustTrivial•20m ago•0 comments

Stats from 30K AI debates: Opus 4.7 is the most influential model

https://opper.ai/ai-roundtable/stats
5•felix089•20m ago•1 comments

How to Build an ML Framework in Rust, from Scratch, in a Weekend

https://www.erikkaum.com/blog/zml/index.html
1•tosh•22m ago•0 comments

NASA Says Farewell to Maven Mars Mission

https://www.nasa.gov/news-release/nasa-says-farewell-to-maven-mars-mission-hosts-media-call-today/
2•ironyman•22m ago•1 comments

Why open standards matter for AI infrastructure

https://openenvelope.org/writing/open-standards-ai-infrastructure/
2•ashconway•23m ago•0 comments

Compiling Zig to RISC-V

https://www.erikkaum.com/blog/advent-05/index.html
3•tosh•24m ago•0 comments

Counterfeit G.Skill and V-Color DDR5 modules hit Chinese marketplaces

https://www.tomshardware.com/pc-components/dram/counterfeit-g-skill-and-v-color-ddr5-modules-hit-...
3•speckx•25m ago•0 comments

The Public Should Own Half of the Big A.I. Companies

https://www.sanders.senate.gov/op-eds/the-public-should-own-half-of-the-big-a-i-companies/
22•droidjj•25m ago•5 comments

If AI Data Centers Are So Great, Why Are They Being Built in Secret?

https://www.thebrockovichreport.com/p/if-data-centers-are-so-great-why
4•thisislife2•27m ago•0 comments

Overwhelmed by AI Cost Management? The Tokenomics Foundation Can Help

https://techstrong.ai/articles/overwhelmed-by-ai-cost-management-the-tokenomics-foundation-can-help/
1•CrankyBear•28m ago•0 comments