CVE-2026-34430
Deerflow ≤ 2026-03-29
Raw vector
CVSS:4.0/AV:N/AC:L/AT:N/PR:N/UI:A/VC:H/VI:H/VA:H/SC:N/SI:N/SA:N/E:X/CR:X/IR:X/AR:X/MAV:X/MAC:X/MAT:X/MPR:X/MUI:X/MVC:X/MVI:X/MVA:X/MSC:X/MSI:X/MSA:X/S:X/AU:X/R:X/V:X/RE:X/U:XSummary
CVE-2026-34430 is a high-severity Incomplete List of Disallowed Inputs (CWE-184) vulnerability in Deerflow Deerflow. Its CVSS base score is 8.6 (High).
Operationally, exploitation aligns with the MITRE ATT&CK technique Exploit Public-Facing Application (T1190); ranked at the 33th percentile by exploit likelihood (below the median); it is not currently listed in the CISA KEV catalog; a public proof-of-concept is referenced.
The strongest mitigations our analysis identified map to SI-10 (Information Input Validation) — see the control section below for these in your framework.
Deeper analysis AI-assisted summary
Synthesised by an AI model from the NVD description and linked references — a reading aid, not an authoritative source.
ByteDance Deer-Flow versions prior to commit 92c7a20 are affected by CVE-2026-34430, a sandbox escape vulnerability in the bash tool handling mechanism. The flaw stems from incomplete modeling of shell semantics, particularly regex-based validation that fails to account for shell features like directory changes (e.g., cd) and relative paths. This allows attackers to bypass sandbox boundaries, enabling reads and modifications of files outside the restricted environment, as well as arbitrary command execution on the host system through subprocess invocation with shell interpretation enabled. The vulnerability is rated 8.8 on the CVSS 3.1 scale (AV:N/AC:L/PR:N/UI:R/S:U/C:H/I:H/A:H) and is associated with CWE-184.
Attackers can exploit this vulnerability remotely over the network with low complexity and no required privileges, though user interaction is necessary, such as tricking a user into executing a malicious input within the Deer-Flow environment. Successful exploitation grants full arbitrary command execution on the host, compromising confidentiality, integrity, and availability with high impact, as attackers can read sensitive files, modify system data, and run unauthorized processes outside the sandbox.
Mitigation is available through updating to commit 92c7a20 or later in the ByteDance Deer-Flow repository, as detailed in the associated GitHub commit (https://github.com/bytedance/deer-flow/commit/92c7a20cb74addc3038d2131da78f2e239ef542e) and pull request #1547 (https://github.com/bytedance/deer-flow/pull/1547). Additional guidance on the vulnerability and remediation is provided in the VulnCheck advisory (https://www.vulncheck.com/advisories/bytedance-deerflow-localsandboxprovider-host-bash-escape).
EU & UK References
- 🇪🇺 ENISA EUVD: EUVD-2026-17903
Vulnerability Data
ByteDance DeerFlow versions prior to commit 92c7a20 contain a sandbox escape vulnerability in bash tool handling that allows attackers to execute arbitrary commands on the host system by bypassing regex-based validation using shell features such as directory changes and relative…
more
paths. Attackers can exploit the incomplete shell semantics modeling to read and modify files outside the sandbox boundary and achieve arbitrary command execution through subprocess invocation with shell interpretation enabled.
- CWE(s)
Related Threats
MITRE ATT&CK Enterprise Techniques
CVEs Like This One
Affected Assets
Mitigating Controls
Control response
—
—
- 2 hardening rules · 1 OS baseline
V3.5.2V4.4.2V16.2.5
Mitigating Controls (NIST 800-53 r5) AI
SI-10 requires validity checks on inputs, which structurally replaces incomplete deny-lists with complete allow-list or sanitization logic.
Mitigating Controls (NIST CSF 2.0) AI
Derived directly from the weakness types (CWEs) cited in the NVD entry via our AI-authored CWE→CSF cross-walk (authority under review) — links open the control.
Secure SDLC practices directly require complete, positive input validation instead of incomplete denylists.
Mitigating Controls (ISO/IEC 27001:2022 Annex A) AI
Derived directly from the weakness types (CWEs) cited in the NVD entry via our AI-authored CWE→ISO cross-walk (authority under review) — links open the control.
Security testing can discover missing input checks, but does not prevent the weakness during development.
Application security requirements can mandate complete input validation rules, but the control itself does not prescribe how to build those rules.
Secure architecture principles include robust input validation design, yet the control is broader than this single weakness.
Secure coding standards directly require exhaustive allow-lists or complete deny-lists for inputs, addressing the root cause of incomplete disallowed-input lists.