AI Threat Watch

Case xai-api-key-github-leakModel & data leakUnited States

xAI employee leaks API key on GitHub exposing 60 private Grok models for two months

A key committed to a public GitHub repo by an xAI staffer worked against at least 60 fine-tuned and unreleased models, including grok-2.5V and models apparently built on SpaceX and Tesla data. It stayed valid two months.

HighAttack on AIStatus: ResolvedVerified case file

Open in the interactive mapCite this case

Reported
May 1, 2025
Location
United States
Target
xAI private and unreleased Grok models
Severity
High
Triggered by
xAI employee (accidental)
Actor type
Insider
Reported loss
None reported
CVE
None

What happened

A key committed to a public GitHub repo by an xAI staffer worked against at least 60 fine-tuned and unreleased models, including grok-2.5V and models apparently built on SpaceX and Tesla data. It stayed valid two months.

Impact: Two months of potential unauthorised access to private LLMs; no confirmed misuse

Who triggered it

Actor: xAI employee (accidental)

Attribution: Found by consultant Philippe Caturegli and GitGuardian

Motive: None (accidental commit)

How it happened

Attack vector: Secret committed to a public code repository

Root cause: No secret scanning or key rotation on employee repositories

Timeline

  1. GitGuardian alerts the employee
  2. GitGuardian alerts xAI security; key still valid
  3. Repository removed after reporting

Response

Repository taken down; xAI did not publicly comment

Current status: Resolved

Sources (open in a new tab)

Cite this case

xAI employee leaks API key on GitHub exposing 60 private Grok models for two months. AI Threat Watch, case xai-api-key-github-leak, May 1, 2025. https://aithreat.io/cases/xai-api-key-github-leak/

Permanent link: https://aithreat.io/cases/xai-api-key-github-leak/

This case file in other languages: EnglishFrançaisEspañolالعربية中文