Disclaimer: This content was generated using AI. While we strive for accuracy, please verify critical information with official sources.
Sources: National Vulnerability Database, CISA KEV, Exploit-DB
CVE-2026-114198 Overview
CVE-2026-114198 is an insecure deserialization vulnerability [CWE-502] in Mandiant FLARE floss releases before version 3.2.0. The flaw allows arbitrary code execution when an analyst loads a crafted .viv or .exe.viv workspace file through the floss utility. Because floss is widely used by malware analysts to extract obfuscated strings from suspicious binaries, the vulnerability turns the analysis tool itself into an attack surface. A malicious workspace shipped alongside a sample can execute attacker-controlled code in the analyst's environment.
Critical Impact
A crafted .viv workspace supplied to floss triggers Python deserialization, giving attackers arbitrary code execution with the privileges of the analyst running the tool.
Affected Products
- Mandiant FLARE floss releases prior to 3.2.0
- Analysis environments ingesting third-party .viv or .exe.viv workspace files
- Automated malware triage pipelines that invoke floss on untrusted inputs
Discovery Timeline
- 2026-10-21 - CVE-2026-114198 published to the National Vulnerability Database
- 2026-10-22 - NVD record last modified
Technical Details for CVE-2026-114198
Vulnerability Analysis
FLOSS (FLARE Obfuscated String Solver) is a static analysis tool that extracts strings from malware samples. It persists analysis state in Vivisect workspace files with .viv or .exe.viv extensions. These workspaces are loaded via Python's deserialization routines, which reconstruct arbitrary objects from the file's byte stream.
The vulnerability stems from loading workspace files without validating their origin or integrity. An attacker who controls the contents of a .viv file can embed malicious serialized objects. When floss deserializes the workspace, it instantiates those objects and triggers attacker-defined callbacks. Execution occurs within the analyst's Python interpreter, inheriting its file system access and network connectivity.
Root Cause
The root cause is unsafe deserialization of untrusted data [CWE-502]. Workspace files are treated as trusted internal state rather than user-controlled input. Python serialization formats do not enforce type restrictions during object reconstruction, allowing arbitrary class instantiation and method invocation via well-known gadget chains.
Attack Vector
Exploitation requires local attack access and user interaction. An attacker crafts a malicious .viv or .exe.viv file and delivers it alongside a malware sample, in a shared threat intelligence feed, or through a compromised repository. When an analyst runs floss against the crafted workspace, deserialization executes the embedded payload. See the Mandiant Vulnerability Disclosure MNDT-2026-0036 for the vendor's technical description.
// No verified public exploit code is available.
// The attack chain follows the standard Python deserialization pattern:
// crafted .viv workspace -> floss load routine -> object reconstruction -> code execution.
Stop Remote Code Execution Before It Runs
Singularity™ Endpoint detects and blocks exploit attempts at machine speed, including zero-days and novel attack techniques.
Stop Exploit AttemptsDetection Methods for CVE-2026-114198
Indicators of Compromise
- Unexpected child processes spawned by python, floss, or Vivisect interpreters during malware triage
- .viv or .exe.viv files received from external sources or bundled with suspicious samples
- Outbound network connections initiated by analysis workstations immediately after loading a workspace file
- Writes to persistence locations (cron, systemd units, startup folders) from the floss process tree
Detection Strategies
- Hunt for floss or vivisect processes invoking shell interpreters, curl, wget, or scripting runtimes
- Alert on .viv workspace files whose SHA-256 hashes do not match internally generated analyst artifacts
- Correlate file reads of .viv extensions with subsequent anomalous process creation on sandbox hosts
Monitoring Recommendations
- Enable process lineage telemetry on malware analysis workstations and sandbox virtual machines
- Capture command-line arguments passed to floss to identify loads of externally sourced workspaces
- Forward endpoint and sandbox telemetry to a central data lake for retroactive hunting against newly identified malicious .viv hashes
How to Mitigate CVE-2026-114198
Immediate Actions Required
- Upgrade Mandiant FLARE floss to version 3.2.0 or later across all analyst workstations and automated pipelines
- Audit shared repositories and threat intelligence feeds for untrusted .viv and .exe.viv files and remove them
- Restrict floss execution to isolated, non-privileged sandbox environments without access to sensitive internal networks
Patch Information
Mandiant resolved the vulnerability in FLARE floss release 3.2.0. Refer to the Mandiant Vulnerability Disclosure MNDT-2026-0036 for release details and upgrade guidance.
Workarounds
- Only load .viv workspaces that were generated locally by trusted analysts
- Run floss inside ephemeral containers or virtual machines with no outbound network access
- Treat all third-party workspace files as untrusted input and regenerate them from the original sample before analysis
# Upgrade floss to the fixed release
pip install --upgrade 'flare-floss>=3.2.0'
# Verify the installed version
floss --version
Disclaimer: This content was generated using AI. While we strive for accuracy, please verify critical information with official sources.
Sources: National Vulnerability Database, CISA KEV, Exploit-DB

