frontpage.
newsnewestaskshowjobs

Made with ♥ by @iamnishanth

Open Source @Github

fp.

Open in hackernews

Show HN: BreakMyAgent – Open-source red-teaming sandbox for LLM system prompts

2•breakmyagent•1h ago
As a developer, I got tired of manually testing my AI agents and chatbots against the same prompt injections and jailbreaks every time I tweaked a system prompt. Our QA team was struggling with the exact same bottleneck, so I built BreakMyAgent.

It’s an open-source sandbox that runs an automated barrage of standard exploits against your target LLM to see if it leaks data or ignores core instructions.

How it works under the hood: - The UI is built with Streamlit, backend is FastAPI, and dependency management is handled by `uv`. - You paste your system prompt and hit run. It fires 12 baseline attack vectors (Direct leaks, XSS payloads, Context overflows, etc.) concurrently. - The core mechanic is "LLM-as-a-Judge". It uses a hardcoded `gpt-4.1-mini` with strict alignment rules to systematically evaluate the target's responses. - It supports OpenAI, Anthropic, and a solid list of open-weight models via OpenRouter (including DeepSeek V3/R1, Qwen 2.5, and Llama 3.3).

There is a hosted free version if you want to play with it immediately (I capped it at 15 requests/IP to survive the launch), but the entire tool is open-source and takes 30 seconds to spin up locally with Docker or `uv`.

Repo: https://github.com/BreakMyAgent/breakmyagent-os Live demo: https://breakmyagent.dev

Next on the roadmap: I'm building a dedicated CLI/GitHub Action so teams can drop this into their own CI/CD pipelines to block prompt regressions. I'm also developing a PoC for multi-turn agentic fuzzing and expanding the payload database for complex tool-spoofing.

I’d love to hear your feedback! What other test configurations (besides temperature and response format) do you think are essential for a tool like this? Also open to any feedback on the architecture, the judge prompt, or specific zero-day vectors you'd like to see included in the public database.

Five ways to spot when a paper is a fraud

https://www.nature.com/articles/d41586-026-00569-x
1•bookofjoe•31s ago•1 comments

Riot's New Fighting Game Is Imploding as It Lays Off 80 Developers

https://kotaku.com/2xko-layoffs-league-legends-riot-update-2000666998
1•PaulHoule•40s ago•0 comments

Snipit – A lightweight CLI to save and search code snippets locally

https://github.com/fouadbuilds/snipit
1•fouaden•1m ago•1 comments

Show HN: MVAR – Deterministic sink enforcement for AI agent

https://github.com/mvar-security/mvar
1•ShawnC21•2m ago•0 comments

Are you sure you're burning enough tokens?

https://www.openbattle.club/
1•nunojay•2m ago•0 comments

Every AI code review vendor benchmarks itself, and wins

https://deepsource.com/blog/notes-on-ai-code-review-benchmarks
1•dolftax•5m ago•0 comments

CesiumAstro Announces Acquisition of Vidrovr

https://finance.yahoo.com/news/cesiumastro-announces-acquisition-vidrovr-enhance-130000040.html
1•danielmorozoff•6m ago•0 comments

AI Agents Want to Write TypeScript

https://encore.dev/blog/typescript-ai
1•andout_•7m ago•0 comments

History's Best Strategies for Avoiding Being Buried Alive

https://www.atlasobscura.com/articles/users-guide-to-definitive-death
1•Brajeshwar•7m ago•0 comments

AI models are being prepared for the physical world

https://www.economist.com/science-and-technology/2026/02/25/ai-models-are-being-prepared-for-the-...
1•Brajeshwar•7m ago•0 comments

One-stop blood tests for multiple types of cancer are increasingly popular

https://www.economist.com/science-and-technology/2026/02/25/one-stop-blood-tests-for-multiple-typ...
1•Brajeshwar•7m ago•0 comments

Unit testing your code's performance, part 2: Testing speed

https://pythonspeed.com/articles/speed-unit-tests/
1•todsacerdoti•7m ago•0 comments

Robert Kaye, MetaBrainz Founder and Executive Director, Has Died

https://blog.metabrainz.org/2026/02/24/robert-kaye/
2•CharlesW•9m ago•0 comments

Cause-specific excess mortality in rural India during Covid-19 pandemic 2020–23

https://bmjopen.bmj.com/content/16/2/e097857
1•Anon84•9m ago•0 comments

Show HN: Multiplayer realtime text-to-website demo (live edits via Sonnet 4.6)

https://textyoursite.com/demo
1•elliotbnvl•10m ago•0 comments

Large language models reflect the ideology of their creators

https://www.nature.com/articles/s44387-025-00048-0
1•geox•10m ago•0 comments

Lofi Car

https://loficar.com
1•kilroy123•13m ago•0 comments

Penguins Are Solar Geoengineers

https://www.governance.fyi/p/all-natural-geoengineering-with-frank-a9d
1•bigbobbeeper•14m ago•0 comments

Show HN: Simple Viewers – Tiny native macOS file viewers

https://www.ryanlitalien.com/simple/
2•ryanlitalien•15m ago•0 comments

Worb: Local open-source wandb-compatible server

https://worb.cloud
1•psarna•16m ago•0 comments

Accenture: You're promoted or fired on using the AI

https://pivot-to-ai.com/2026/02/25/accenture-youre-promoted-or-fired-on-using-the-ai/
1•ColinWright•17m ago•0 comments

US role as global talent hub in doubt amid Donald Trump's visa crackdown

https://www.ft.com/content/c8114fd1-771b-49ac-98c3-a8acf6177626
2•alephnerd•18m ago•2 comments

Do you have to be polite to AI?

https://www.bbc.com/future/article/20260224-the-best-way-to-talk-to-a-chatbot
1•Sikara•19m ago•1 comments

Solving Impossible Problems for Fun and Profit – Dan Gelbart

https://www.youtube.com/watch?v=UTgrWmOk4q8
1•o4c•20m ago•1 comments

Firefox 148 introduces the AI kill switch for people who aren't into LLMs

https://www.xda-developers.com/firefox-148-introduces-the-promised-ai-kill-switch-for-people-who-...
3•randycupertino•21m ago•0 comments

Show HN: I built a 50ms SPF record and Shadow IT scanner

https://spf1.com
2•bwoud•21m ago•3 comments

Show HN: Typed overlay over SQL now supports DuckDB

https://www.datahaskell.org/blog/2026/02/25/beam-duckdb-release.html
1•cosmic_quanta•21m ago•0 comments

Foundation Models SDK for Python Documentation

https://apple.github.io/python-apple-fm-sdk/
1•alexellisuk•21m ago•1 comments

Don't Panic: 'Humanity's Last Exam' Has Begun

https://stories.tamu.edu/news/2026/02/25/dont-panic-humanitys-last-exam-has-begun/
1•thunderbong•22m ago•0 comments

High temperatures affect sex ratios at birth

https://www.ox.ac.uk/news/2026-02-23-new-research-shows-high-temperatures-affect-sex-ratios-birth
1•codingbuddy•22m ago•0 comments