# CVE-2026-108387: radare2 command execution via crafted binary format

Source: https://junglewise.ai/threats/cve-2026-108387-radare2-is-a-unix-like-reverse-engineering-framework-and-command  
Publisher: Junglewise Threat Intelligence, https://junglewise.ai/threats  
Published: 2026-11-04. Updated: 2026-11-04  
License: CC BY 4.0 with a link back

CVE-2026-108387 · Severity: info · Published 2026-11-04

radare2 is a reverse engineering tool used to analyze and debug binary programs. When opening a specially crafted Pebble or Sega Mega Drive binary file, an attacker can inject malicious commands that radare2 executes with the same privileges as the user running the tool, potentially leading to unauthorized code execution or system compromise.

Technologies: [Radare2](https://junglewise.ai/threats/technologies/radare2). Vendors: [Radare](https://junglewise.ai/threats/vendors/radare).

## Quick answers

**Is CVE-2026-108387 exploited in the wild?** No. Junglewise Threat Intelligence has no report of CVE-2026-108387 being exploited in the wild.

**How severe is CVE-2026-108387?** Informational severity, published 4 November 2026.

**Which products and versions are affected?** Radare2 before 6.2.2.

**Is there a fix for CVE-2026-108387?** Yes: fixed in 6.2.2. Junglewise finds every place you run it, drafts the upgrade and applies it once you approve.

**How is CVE-2026-108387 attacked?** It is attacked with local access, in a software library.

**Am I affected by CVE-2026-108387?** Check whether you run Radare2 at an affected version. Junglewise does that check for you around the clock: connect your code, cloud and domains, and it matches new vulnerabilities like this one against what you actually run and tells you when one reaches you.

## How to fix

- Fixed in 6.2.2

## Technical details

A command injection vulnerability exists in radare2's binary format plugins where attacker-controlled data from Pebble and Sega Mega Drive binaries can reach the RBinInfo->bclass field. By injecting a dbg:// URI into bclass, the tool automatically reopens the input through the debug I/O plugin and passes attacker-controlled arguments to execvp(), bypassing the need for explicit debugging commands. This allows arbitrary command execution in the context of the radare2 process.

## Affected products

- radare radare2 before 6.2.2

## Timeline

- 2026-11-04: disclosed
- 2026-11-04: patched

## References

- [https://github.com/radareorg/radare2/commit/59899e6b8b18772ce9460b4caa2ced7f257fae76](https://github.com/radareorg/radare2/commit/59899e6b8b18772ce9460b4caa2ced7f257fae76)
- [https://github.com/radareorg/radare2/commit/c9de154cafd788d66535435e43ad6a54aed7ea2a](https://github.com/radareorg/radare2/commit/c9de154cafd788d66535435e43ad6a54aed7ea2a)

## Related threats

- [radare2 infinite loop in NE relocation parser](https://junglewise.ai/threats/cve-2026-81885-radare2-is-a-unix-like-reverse-engineering-framework-and-command)
- [radare2 is a UNIX-like reverse engineering framework and command-line toolset. Prior to 6.2.0, radare2's binary…](https://junglewise.ai/threats/cve-2026-81882-radare2-is-a-unix-like-reverse-engineering-framework-and-command)
- [radare2 is a UNIX-like reverse engineering framework and command-line toolset. Prior to 6.2.0, radare2's Mach-O Swift…](https://junglewise.ai/threats/cve-2026-81881-radare2-is-a-unix-like-reverse-engineering-framework-and-command)
- [radare2 PEF loader denial of service](https://junglewise.ai/threats/cve-2026-81880-radare2-is-a-unix-like-reverse-engineering-framework-and-command)
- [radare2 heap out-of-bounds read in ELF PN_XNUM handling](https://junglewise.ai/threats/cve-2026-81879-radare2-is-a-unix-like-reverse-engineering-framework-and-command)