Bytebot CVE-2026-30631
CRITICALSeverity by source
AV:N/AC:L/PR:N/UI:N/S:U/C:H/I:H/A:H
Unauthenticated command injection reachable via the agent's network API yields code execution with full host impact, matching AV:N/AC:L/PR:N and C:H/I:H/A:H; assumes the control API is exposed.
Primary rating from Vendor (mitre).
CVSS VectorVendor: mitre
Lifecycle Timeline
5DescriptionCVE.org
An issue was discovered in bytebot-ai in commit 3d37894ce07ef8d8b40adc7fd309ad96c2a71313 (2025-09-11) allowing attackers to execute arbitrary code via crafted path to computer_write_file.
AnalysisAI
Remote arbitrary code execution in Bytebot (bytebot-ai self-hosted AI desktop agent, commit 3d37894 dated 2025-09-11) allows attackers to run OS commands by supplying a crafted path to the agent's computer_write_file action. The flaw is classed as OS command injection (CWE-78) and carries a CVSS 9.8, but real-world exploitation probability is currently low (EPSS 0.17%, 7th percentile) and it is not on CISA KEV. A public gist referenced by the CVE appears to document the issue and may contain proof-of-concept details, though POC status is not explicitly confirmed in the source data.
Technical ContextAI
Bytebot is an open-source, self-hosted AI 'computer-use' agent that exposes a control API letting an LLM-driven agent perform desktop actions such as typing, clicking, and file manipulation. One such action, computer_write_file, accepts a file path parameter; per the description a crafted path is passed into an OS-level operation without adequate sanitization, matching CWE-78 (Improper Neutralization of Special Elements used in an OS Command). Rather than treating the path purely as a filesystem argument, attacker-controlled path input reaches a shell/command context, enabling command injection and arbitrary code execution. The CPE data is uninformative (cpe:2.3:a:n/a:n/a - a placeholder), so exact affected-product enumeration relies on the GitHub repository and the referenced fix commit rather than NVD CPE.
RemediationAI
Upstream fix available (commit); a released patched version is not independently confirmed - update Bytebot to a build that includes commit 3d37894ce07ef8d8b40adc7fd309ad96c2a71313 (or later main) which addresses the computer_write_file path handling. Until you can deploy the fixed code, restrict exposure of the Bytebot agent's control API: do not bind it to public/untrusted interfaces, place it behind network ACLs or a firewall so only trusted orchestrators can reach the computer_write_file endpoint, and add authentication/reverse-proxy auth in front of the API if none exists (trade-off: may break any client that assumes an open local API). If file-write functionality is not required for your workflow, disable or gate the computer_write_file action (trade-off: loses agent file-writing capability). Run the agent in a locked-down container/VM with least privilege so a successful injection is contained. Note the referenced commit also bumps Prisma dependencies (@prisma/client 6.6.0→6.16.1), so review the full diff before cherry-picking.
LibreNMS before 24.10.0 allows a remote attacker to execute arbitrary code via OS command injection involving AboutContr
Reachable assertion in Open5GS up to version 2.7.6 allows remote unauthenticated attackers to crash the Diameter CER han
A Command Injection issue in the payload build page in BYOB (Build Your Own Botnet) 2.0 allows attackers to execute arbi
Denial of service in Open5GS 2.7.6 and earlier allows network-adjacent attackers with low-privilege Diameter Gx peer acc
Denial of service in Open5GS HSS (Home Subscriber Server) versions 2.7.0 through 2.7.6 is triggered remotely by manipula
Remote command injection in Control Web Panel allows unauthenticated attackers to execute arbitrary OS commands as root
LINQPad before 5.52.01 Pro edition is vulnerable to Unsafe Deserialization in LINQPad.AutoRefManager::PopulateFromCache(
Remote code execution in dash-uploader (Python package for Plotly Dash) versions 0.1.0 through 0.7.0a2 allows unauthenti
Command injection in LiteLLM 1.18.10 lets attackers who can configure MCP (Model Context Protocol) servers supply arbitr
SQL injection in AzuraCast's backup restore flow (versions up to and including 0.23.7) lets an attacker achieve privileg
Arbitrary file read in Yu AI Code Mother v4.3 lets anonymous, remote attackers retrieve files outside the intended previ
Remote code execution in APScheduler (all versions through 3.10.x and 4.0.0a5) is achievable when applications deseriali
Same weakness CWE-78 – OS Command Injection
View allShare
External POC / Exploit Code
Leaving vuln.today