Fireworks AI: eval-job environment variables and API keys exposed via stolen internal credentials

Severity: high · Patch status: unknown · Since 2026-10-09

Fireworks AI says an unauthorized third party used internal Fireworks credentials to read environment variables attached to evaluation jobs in a limited number of customer accounts; API keys and tokens stored as secrets were likely captured. Fireworks found the activity on Oct 6, notified customers from Oct 7, rotated employee API keys on Oct 9 and revoked more user keys and service-account tokens on Oct 10. It reports no evidence that customer models, datasets, training data, inference traffic or personal information were accessed, but has not confirmed or ruled out misuse of the exposed third-party credentials. Affected customers must rotate every credential listed in their notice at the issuing provider (deleting the stored secret does not invalidate it). Wiz counts it among at least nine AI developer platforms hit in October 2026.

What to do

If you received a Fireworks notice, rotate every third-party credential it lists at the issuing provider, then update apps and Fireworks secrets; deleting the stored secret does not invalidate it
Create new Fireworks API keys to replace revoked ones and update applications
Review logs of the listed third-party services for unauthorized reads, downloads or changes
Look for unrecognized API keys, users, evaluators, training jobs and datasets and report them to security@fireworks.ai
Audit what secrets your eval jobs and other hosted jobs receive; keep high-value keys out of them

Sources

All live incidents on Threat Watch

Threat Watch– active

All incidents

More on Threat Watch

More Fireworks AI incidents

Uber: driver personal data accessed via compromised credentials at a service provider (no CVE) Medium · Breach
Notion: AI chat transcripts in Fireworks AI breach dataset, to be treated as exposed (LLMJacking-linked) Medium · Breach
Fireworks AI: second breach via third-party cloud provider exposes customer billing data and model assets High · Breach
Black October: stolen platform-stored AI keys fuel Bedrock LLMjacking across at least nine platforms High · Breach · Closed
Claude Code extension (Cursor marketplace): tampered 2.1.312/2.1.313 plants SSH backdoor (no CVE) High · Supply chain · Closed

Browse

Helix Cyber

Is your stack exposed? Get it tested.

The incidents above hit someone else first. A Helix pentest finds what an attacker would use against your web apps, APIs and cloud. A second agent verifies each finding, and retests run from the same portal.

Get a pentest How it works

Agent Virtualization

Enterprise Coding Agents

Private AI Platform

Digital Sovereignty

Follow the Sun

The Speed Advantage

Financial Services

GPU Cloud & Neoclouds

Public Sector & Defence

More on Threat Watch

More Fireworks AI incidents

Browse

Is your stack exposed? Get it tested.