CVE-2026-129877: DeepAudit exposes LLM API keys and Git tokens to unauthenticated attackers

DeepAudit versions 2.0.0-beta.1 through 3.8.0 expose server secrets, including LLM provider API keys and GitHub or GitLab tokens, through an unauthenticated endpoint.

Christopher Ort
By Christopher Ort · 3 min read
Share
A server room with exposed cables and open cabinet doors, symbolizing leaked credentials and unsecured infrastructure
Image: AI-generated, i10X
On this page
  1. Quick Take
  2. In plain English
  3. What happened
  4. What you should do
  5. What it means
  6. Sources
i10X Quick Take

Quick Take

  1. What: DeepAudit 2.0.0-beta.1 through 3.8.0 exposes LLM API keys and Git tokens via an unauthenticated endpoint.
  2. Who is affected: Any organization running DeepAudit versions 2.0.0-beta.1 through 3.8.0 with configured LLM provider keys or GitHub/GitLab tokens.
  3. CVE: CVE-2026-129877, CVSS 7.5 (high severity).
  4. What to do: Rotate all LLM provider API keys and GitHub/GitLab tokens immediately, then upgrade to a patched version when available.

In plain English

DeepAudit is a developer tool that uses AI to audit code. A security flaw in versions 2.0.0-beta.1 through 3.8.0 lets anyone on the internet read the server’s secret keys without logging in. These secrets include API keys for AI services like OpenAI, Claude, and Gemini, plus tokens for GitHub and GitLab. An attacker can use these keys to run up bills on your AI accounts or access your private code repositories.

What you should do: If you run DeepAudit, you need to change all those keys and tokens right away. If you don’t use DeepAudit, you can ignore this.

What happened

DeepAudit versions 2.0.0-beta.1 through 3.8.0 contain an information disclosure vulnerability tracked as CVE-2026-129877. The vulnerability exists in the /api/v1/config/defaults endpoint, which returns server configuration data without requiring authentication.

According to the vulnerability report, an attacker can send a GET request to this endpoint and retrieve the operator’s LLM provider API keys and GitHub or GitLab tokens. The vulnerable code is present in the backend/app/api/v1/endpoints/config.py file at lines 109 through 154 in commit 9aa46568f20b2ab2dbc1878a4479392012c1230a.

The exposed secrets include API keys for multiple LLM providers (OpenAI, Claude, Gemini, Qwen, DeepSeek, Zhipu, Moonshot, Baidu, Minimax, and Doubao) as well as GitHub and GitLab access tokens. Once obtained, these credentials can be used to abuse the associated services and access repositories that the tokens have permission to reach.

The vulnerability was discovered by Wu Yulun and disclosed on November 6, 2026. The National Vulnerability Database assigned it a CVSS score of 7.5 (high severity), while VulnCheck assigned a CVSS v4 score of 8.7.

What you should do

If you operate a DeepAudit instance running versions 2.0.0-beta.1 through 3.8.0, you should immediately rotate all LLM provider API keys and GitHub or GitLab tokens that were configured in the system. Assume these credentials have been compromised.

The sources do not state whether a patched version is available. Monitor the DeepAudit GitHub repository for security updates and apply them as soon as they are released.

Review access logs for the /api/v1/config/defaults endpoint to determine whether unauthorized parties accessed the configuration data.

What it means

This vulnerability represents a critical failure in authentication controls for sensitive configuration data. The exposure of LLM provider API keys can lead to unauthorized usage charges and potential data exfiltration if the keys have access to fine-tuned models or stored prompts. The exposure of GitHub or GitLab tokens can grant attackers access to private source code repositories, potentially leading to intellectual property theft or supply chain attacks.

The vulnerability affects all DeepAudit installations from version 2.0.0-beta.1 through 3.8.0, a range that includes both beta and stable releases. Organizations using DeepAudit for code auditing should treat this as a high-priority security incident requiring immediate credential rotation and system review.

Leaked AI provider keys are a recurring pattern. The Railway breach led to $740,000 in unauthorized charges, and the Hugo Blog AI Proxy exposed API keys without authentication. New cases are tracked in our AI Security hub.

Sources

  • NVD CVE-2026-129877
  • DeepAudit vulnerable code on GitHub
  • DeepAudit Vulnerability Report
  • VulnCheck Advisory
Christopher Ort

Christopher Ort

Christopher Ort handles SEO, AI agent workflows and PR & partnerships at i10x.ai, the all-in-one AI workspace built by i10X Global Pte. Ltd. in Singapore. He is a graduate of Friedrich-Alexander-Universität Erlangen-Nürnberg (FAU). He is the author of “The Style Penalty”, a study published on SSRN that analyses 1,576 evaluation decisions by four frontier AI models to show how AI-written CVs fare in automated hiring screens; the research has been covered by heise online and t3n.

Editor's picks

  1. CVE-2026-129877: DeepAudit exposes LLM API keys and Git tokens to unauthenticated attackers
  2. CVE-2026-129875: REBUILD 4.5.0–4.5.4 password hash disclosure via MCP QueryRecords
  3. South Korea Weighs Nuclear Power to Meet AI Electricity Demand
  4. UK and US Partner to Accelerate Nuclear Fusion with AI
  5. Texas and New Mexico Data Center Emissions Rise With AI Buildout