CVE-2025-58357

9.6 CRITICAL

📋 TL;DR

CVE-2025-58357 is a content injection vulnerability in 5ire AI assistant that allows attackers to inject malicious content through chat page script gadgets. This affects all users of 5ire version 0.13.2 who interact with malicious prompts, compromised MCP servers, or exploited tool integrations.

💻 Affected Systems

Products:
  • 5ire AI Assistant
Versions: Version 0.13.2
Operating Systems: Windows, Linux, macOS
Default Config Vulnerable: ⚠️ Yes
Notes: All installations of version 0.13.2 are vulnerable regardless of configuration.

📦 What is this software?

⚠️ Risk & Real-World Impact

🔴

Worst Case

Complete system compromise through arbitrary code execution, data theft, and lateral movement within the network.

🟠

Likely Case

Session hijacking, credential theft, and unauthorized access to sensitive information through injected malicious content.

🟢

If Mitigated

Limited impact with proper input validation and output encoding, potentially only cosmetic content injection.

🌐 Internet-Facing: HIGH - The vulnerability can be triggered through external malicious prompts and compromised servers.
🏢 Internal Only: MEDIUM - Internal users could be targeted through malicious internal prompts or compromised internal servers.

🎯 Exploit Status

Public PoC: ✅ No
Weaponized: UNKNOWN
Unauthenticated Exploit: ⚠️ Yes
Complexity: MEDIUM

Exploitation requires user interaction with malicious content but no authentication is needed once triggered.

🛠️ Fix & Mitigation

✅ Official Fix

Patch Version: 0.14.0

Vendor Advisory: https://github.com/nanbingxyz/5ire/security/advisories/GHSA-8527-3cch-95gf

Restart Required: Yes

Instructions:

1. Download version 0.14.0 from GitHub releases. 2. Uninstall version 0.13.2. 3. Install version 0.14.0. 4. Restart the application.

🔧 Temporary Workarounds

Disable External MCP Servers

all

Prevent connection to potentially compromised MCP servers

Configure 5ire to only use trusted, internal MCP servers

Restrict Tool Integrations

all

Limit which tools can be integrated to reduce attack surface

Review and disable unnecessary tool integrations in 5ire settings

🧯 If You Can't Patch

  • Isolate 5ire instances on separate network segments
  • Implement strict input validation and output encoding at the network perimeter

🔍 How to Verify

Check if Vulnerable:

Check 5ire version in application settings or via '5ire --version' command

Check Version:

5ire --version

Verify Fix Applied:

Confirm version is 0.14.0 or higher and test chat functionality with known safe prompts

📡 Detection & Monitoring

Log Indicators:

  • Unusual chat patterns
  • Unexpected script execution
  • Suspicious MCP server connections

Network Indicators:

  • Unexpected outbound connections from 5ire
  • Suspicious payloads in chat traffic

SIEM Query:

source="5ire" AND (event="script_execution" OR event="external_connection")

🔗 References

📤 Share & Export