LobeChat ReDoS via Malicious GitHub Repository URL Path

Executive Summary

CVE-2026-58578 is a Regular Expression Denial of Service (ReDoS) vulnerability in the LobeHub application, specifically in the GitHub skill import feature. The issue occurs because a user-controlled GitHub repository URL path segment is directly inserted into a dynamically constructed regular expression without escaping special regex characters. This allows an authenticated attacker to craft malicious patterns, such as catastrophic-backtracking regex patterns like `(a+)+`, which cause the regular expression engine to perform excessive backtracking.

When the server processes the repository archive, it synchronously executes this regex against archive entries, blocking the Node.js event loop for tens of seconds per request. This makes the server unresponsive to all other users during that time. Additionally, invalid regex patterns can cause syntax errors resulting in HTTP 500 errors.

The vulnerability is located in the `findSkillMd` function, where the `basePath` from the GitHub URL is used to create the regex. The fix replaces the dynamic regex with safer string matching methods, preventing user input from reaching the RegExp compiler.

Useful 0 Not Useful 0

Detection Guidance

This vulnerability manifests as a server-side delay or denial of service when processing GitHub repository URLs containing malicious regex patterns in the skill import feature. Detection involves monitoring for unusually long response times or HTTP 500 errors from the `agentSkills.importFromGitHub` endpoint.

Specifically, look for requests with GitHub URL paths containing regex metacharacters such as `(a+)+` or invalid patterns like `[invalid` that cause catastrophic backtracking or syntax errors.

Since the issue blocks the Node.js event loop for tens of seconds, you can detect it by observing high CPU usage or event loop delays during skill import requests.

Suggested commands to detect this on a system running the vulnerable LobeHub server include:

• Use network monitoring tools (e.g., tcpdump, Wireshark) to capture requests to the skill import endpoint and inspect URL paths for suspicious regex patterns.
• Monitor Node.js event loop delays using tools like `clinic doctor` or `node --trace-event` to identify blocking caused by regex processing.
• Check server logs for HTTP 500 errors related to invalid regex patterns in GitHub URL paths.
• Example command to search logs for suspicious patterns: `grep -E '\(a\+\)\+|\[invalid' /path/to/lobehub/logs/*`

Useful 0 Not Useful 0

Mitigation Strategies

The primary mitigation is to upgrade LobeHub to version 2.2.10-canary.15 or later, where the vulnerability has been fixed by replacing the vulnerable dynamic regular expression with safe string matching methods.

Until the upgrade can be applied, consider restricting or monitoring authenticated users' ability to import skills from GitHub URLs, especially those containing suspicious regex metacharacters.

Additionally, implement rate limiting or request timeouts on the skill import endpoint to reduce the impact of potential ReDoS attacks.

Review server logs for signs of exploitation attempts and block or throttle offending IP addresses or user accounts.

Backup your data before applying any updates or changes.

Useful 0 Not Useful 0

Impact Analysis

This vulnerability can cause a denial of service condition on the LobeHub server by blocking the Node.js event loop for tens of seconds per malicious request. During this time, the server becomes unresponsive to all users, effectively denying service.

An authenticated attacker can exploit this by submitting specially crafted GitHub repository URLs containing malicious regex patterns, causing prolonged server unavailability.

Additionally, supplying invalid regex patterns can cause server errors (HTTP 500), further disrupting service.

Useful 0 Not Useful 0