---
title: "Chainalysis breach exposes credentials for all customers — Sekura"
url: https://sekura.ai/blog/breach-chainalysis-2026-11/
description: "An on-path attack using offensive agentic techniques exposed usernames, password hashes, and live tokens for every Chainalysis customer, forcing the firm to take its portal and key products offline."
updated: 2026-11-07
generator: sekura build, scripts/build-markdown.mjs
---
Breach analysis · November 7, 2026

# Chainalysis breach exposes credentials for all customers

An on-path attack using offensive agentic techniques exposed usernames, password hashes, and live tokens for every Chainalysis customer, forcing the firm to take its portal and key products offline.

[All posts →](https://sekura.ai/blog/)

![Sekura — owl-in-shield mark](https://sekura.ai/branding/sekura-owl-v3.png)

An on-path attack using offensive agentic techniques exposed usernames, password hashes, and live tokens for every Chainalysis customer, forcing the firm to take its portal and key products offline.

Hypotheses in, proven findings out: a finding without a working proof never reaches the report. 

Blockchain intelligence firm Chainalysis confirmed last week that an attack on its internal systems exposed usernames, password hashes, and live session tokens for every one of its customers.

The story unfolded in stages. On October 30, Chainalysis issued an initial disclosure describing "limited, non-sensitive" exposure. By November 2, its own investigation found something different: customer information had been accessed. On November 4, the firm updated again, confirming that usernames, password hashes, and live tokens were in the attacker's hands. On November 6, [BleepingComputer reported](https://www.bleepingcomputer.com/news/security/chainalysis-says-breach-exposed-credentials-for-all-customers/) that after detecting further abnormal activity, Chainalysis took its login portal offline along with Reactor, Data Solutions, the KYT UI and API, and its address screening products.

The company described the technique as an "on-path attack" combined with what it called offensive agentic techniques. Live tokens are session credentials that authenticate active requests without passwords. Exposing them is different from exposing password hashes: a cracked hash takes time; a live token works immediately. Chainalysis says it found no evidence that investigative graph data or blockchain attribution data was altered, though "no evidence" and "confirmed absent" are not the same thing. Customers were required to register trusted IP addresses before regaining access, then rotate all passwords and API keys.

Two things are worth pulling out of the reporting.

## What scanners would have missed

Conventional vulnerability scanners would have assessed Chainalysis's login portal for injection flaws, missing HTTP security headers, and known CVEs in its stack. What they would not have done is trace the lifecycle of a live token after it was issued.

This is a consistent gap in the [autonomous penetration testing](https://sekura.ai/learn/what-is-autonomous-penetration-testing/) model scanners try to approximate. A CVSS score measures the theoretical severity of a flaw in isolation. It does not measure whether that flaw, in this environment, in this configuration, would let an on-path adversary collect live tokens at scale. The Chainalysis attack did not require a critical CVE. It required a position in the traffic path and an automated capability to enumerate what was reachable from there.

A DAST scan of the portal surface would have checked for:

1. Session fixation and cookie security flags.
2. Token expiration enforcement on the server side.
3. Brute-force controls at the login boundary.

None of those checks model an adversary who already holds a position between services. The attack surface was not the portal itself. It was the layer between the portal and the internal tools, and that layer is invisible to single-surface scans. The initial Chainalysis disclosure said impact was limited. That assessment was wrong, and it was wrong in a direction that left live tokens in circulation for days.

## What Sekura would have shown

Phase 5 of our pipeline is exploit-chain analysis. It traces multi-step attack paths from initial access to their full downstream impact.

The question here is not "can someone access the login portal?" It is: "if an on-path position is established, what can be collected, and how far does that access extend?" Those are different questions and they require different tooling to answer.

Our exploit-chain agent would have produced a finding showing the path from an on-path position to live token extraction, then traced which internal systems those tokens could reach. It would have shown whether tokens issued for Reactor and the KYT API carry separate scope, and whether expiry logic is enforced consistently across services.

sequenceDiagram participant Attacker participant Proxy as On-Path Position participant Portal as Login Portal participant Token as Token Store participant Reactor as Reactor / KYT API Attacker->>Proxy: Intercepts credential flow Proxy->>Token: Extracts live session token Attacker->>Reactor: Authenticates with live token Reactor-->>Attacker: Returns customer data and org roles 

This is not a claim that we would have prevented this incident. It is a description of what a proof-first finding would have looked like before it: a working chain, with real tokens, against real internal services, showing the actual blast radius rather than a theoretical severity score. That evidence belongs in the design review, not the post-breach review.

## The bigger pattern

The Chainalysis breach is an authentication lifecycle failure. The tokens existed. They were live. They were reachable from the wrong position. The initial exposure assessment was narrow because it examined one surface at a time, not the aggregate state of credentials in motion.

I think this is a 2026 problem in a specific way. As more systems issue short-lived tokens and rely on them for service-to-service trust, the attack surface for on-path collection grows. The token becomes the credential, and the credential is everywhere at once.

The security industry is still applying a snapshot model to a continuous credential surface. That gap is not going to close on its own.

If you want to see what exploit-chain analysis looks like on your authentication layer, [request a POC](https://sekura.ai/enterprise/#poc).

## Related reading

- [Hallmark Salesforce Breach Exposes 1.7 Million RecordsApril 1, 2026](https://sekura.ai/blog/breach-hallmark-cards-2026-04/)
- [Carnival breach exposes 6 million customer recordsApril 8, 2026](https://sekura.ai/blog/breach-carnival-cruise-2026-04/)
- [France Titres breach exposes 11 million citizen recordsApril 22, 2026](https://sekura.ai/blog/breach-france-titres-2026-04/)

[All posts](https://sekura.ai/blog/)
