> ## Content Index
> Fetch the complete content index at: https://www.testingcatalog.com/llms.txt
> Use this file to discover other available public pages before exploring further.

# Fabraix opens public access for Nyx AI security red team agent
- URL: https://www.testingcatalog.com/fabraix-opens-public-access-for-nyx-ai-security-red-team-agent/
- Published: 2026-08-19T18:48:35.000Z
- Updated: 2026-08-19T18:48:35.000Z
- Description: Fabraix opened its second public launch of Nyx, an autonomous red-teaming agent that attacks AI agents to find security holes first.
- Author: Nero Soares
- Tags: Sponsored

Fabraix has opened a second public launch of Nyx, its autonomous red-teaming agent, positioning the product as continuous security testing for teams shipping customer-facing AI. Nyx takes an endpoint, a URL, or a phone number as a target, configures its own test plan from a description, and runs adversarial attacks against chat, voice, browser, and coding agents on a repeating schedule.

![Attack samples](https://storage.ghost.io/c/2a/1b/2a1b1782-8506-4d7d-bf53-ad3fb52e2a0f/content/images/2026/08/image-3.png)

The agent works from a library of more than 10,000 jailbreaks and attack strategies collected from public records, using them as starting points for multi-turn attacks that adapt in real time to how a target defends itself. Testing runs against the agent directly and indirectly through its environment, with payloads placed inside webpages, documents, files, messages, and tool outputs hosted on controlled replicas of SaaS products that Fabraix maintains. Nyx operates as a pure blackbox and requires no source code, model weights, credentials, or network access, with each run kept isolated and ephemeral and never trained on. Every finding arrives with the attack steps, the agent responses, and the failure that resulted, so the same test can be replayed after a release to check whether the fix held. Fabraix reports that Nyx often surfaces a first vulnerability within minutes or hours, and cites a 78 percent attack success rate on AgentHarm, a benchmark for offensive AI security.

Access runs through a CLI installed via npm alongside a REST API, with each audit defined in a YAML file covering target, objective, and budget, and results streaming live during a run. Token-based authentication covers CI pipelines, so audits can sit inside an existing release process instead of arriving as a one-off engagement. Fabraix also maintains a public Playground where the community attempts to break live agents with published system prompts, and publishes research alongside the product, including Adversarial Cost to Exploit, a benchmark that scores AI security by the token spend an attacker needs to breach an agent.

SPONSORED 

Demo booking is available now

[Learn more ](https://fabraix.com/?ref=testingcatalog.com) 

Fabraix was founded in 2026 by Ahmed Aly and Ibrahim Abdu and is backed by Y Combinator in its Summer 2026 batch. Abdu built AI agents at Meta that diagnosed and fixed production errors, after earlier work on compilers and database engines in fintech. Aly led the international payments fraud team at Monzo and was the first data scientist at a Sequoia-backed startup, where he built a fraud engine that grew to process more than a billion dollars in annual B2B transactions. The company reports that Nyx has already found exploitable failures in public-facing agents run by dozens of Fortune 500 companies. The intended audience is security leaders and internal red teams at mid-size and enterprise companies putting agents in front of customers, along with the engineers currently doing that testing by hand. Access is arranged through a demo request, and no public pricing has been listed.