MOOS-IvP StringToIvPFunction() Buffer Overflow
First seen Sep 6, 2026 · Updated Sep 6, 2026 · CVSS 9.8
MOOS-IvP, an autonomous marine vehicle behavior and control framework, contains a critical buffer overflow vulnerability in its IvP function parsing logic (CVE-2026-85438). Attackers who can supply crafted BHV_IPF payloads can trigger out-of-bounds writes leading to memory corruption and potential remote code execution. This affects autonomous vehicle control software rather than typical enterprise AI agent stacks, though similar parsing patterns are common in agent tool pipelines.
Technical Analysis
CVE-2026-85438 (CVSS 9.8) arises in StringToIvPFunction() where dimension, piece, and degree counts parsed from encoded BHV_IPF payloads are used directly as allocation sizes and loop bounds without validation. An attacker can supply a payload with a mismatched dimension count relative to the actual data size, causing the parser to write attacker-controlled double-precision floating point values past the end of the allocated IvPBox weight array. This heap-based buffer overflow can corrupt adjacent memory structures, function pointers, or control data, enabling memory corruption and potential arbitrary code execution on hosts running the MOOS-IvP autonomy stack. Exploitation requires the ability to inject or influence BHV_IPF-encoded messages into the MOOSDB or behavior pipeline, such as through a compromised sensor feed, spoofed inter-vehicle communication, or malicious mission file. While MOOS-IvP is primarily used in unmanned marine vehicle (UMV) autonomy rather than conventional LLM/agent frameworks, any AI agent system that ingests IvP-encoded data as a tool input or bridges to MOOS-IvP for robotic decision-making inherits this deserialization-style risk, underscoring the broader lesson that unvalidated size fields in structured payloads consumed by agent tool-calling pipelines remain a systemic RCE vector.
Affected Systems
MOOS-IvP through version 24.8.1, including all deployments using the StringToIvPFunction() parser for BHV_IPF-encoded messages; unmanned marine/surface/underwater vehicle autonomy stacks; any custom applications or bridges that parse IvP function strings using the vulnerable library code
Indicators of Compromise
- No specific IOCs published; vulnerability is in open-source parsing code (StringToIvPFunction) rather than tied to observed campaign artifacts. Organizations should monitor for anomalous BHV_IPF payload sizes and malformed dimension/piece/degree fields in MOOSDB traffic.
Remediation Steps
- 1
Patch MOOS-IvP
Upgrade to a patched version of MOOS-IvP once released by maintainers; monitor the official MOOS-IvP GitHub repository and NVD for fix availability.
- 2
Input validation at parser boundary
Add strict bounds checking on dimension, piece, and degree fields in StringToIvPFunction() before using them as allocation sizes or loop bounds; reject payloads with inconsistent or oversized values.
- 3
Network/message source restriction
Restrict which hosts, sensors, or vehicles can inject BHV_IPF-encoded messages into the MOOSDB; use authentication and message integrity checks on inter-process and inter-vehicle communications.
- 4
Memory protection hardening
Compile MOOS-IvP binaries with stack canaries, ASLR, DEP/NX, and fortified source functions to reduce exploitability of the overflow.
- 5
Runtime monitoring
Deploy fuzzing and runtime sanitizers (ASan) in test environments to catch malformed IvP function payloads, and add logging/alerting for parser exceptions or crashes in production.
CVE / Advisory IDs
Industries Most Exposed
Respond to this threat
Pro subscribers get a full AI-generated incident-response playbook for this threat — detection, containment, eradication, and recovery steps — plus an unlimited AI Threat Advisor for questions about your environment.