Security readout for executives and security teams
Plain-English summary
A specially crafted TensorFlow Lite model can crash affected TensorFlow processing by triggering a divide-by-zero in the GatherNd operator. The known impact is limited availability disruption, not data theft or privilege escalation. Business urgency is low unless the organization accepts models from untrusted users or partners.
Executive priority
Treat as a low-priority patching item unless untrusted model ingestion exists. For organizations with public or partner model upload workflows, prioritize runtime upgrades and model trust controls to reduce denial-of-service exposure.
Technical view
TensorFlow Lite's reference GatherNd implementation can divide by zero when a crafted model supplies an empty params tensor, making at least one params_shape dimension zero. The advisory maps this to CWE-369 with CVSS 3.1 score 2.5, local attack vector, high complexity, low privileges, and availability-only impact.
Likely exposure
Exposure is most plausible in systems running affected TensorFlow or TFLite versions that load externally supplied models. Risk is lower where model files are internally built, signed, curated, and dependencies are already patched. Check server inference images, developer environments, mobile builds, and edge deployments.
Exploitation context
The provided sources do not show active exploitation, and CISA KEV status is false. Exploitation requires a crafted model and conditions where that model is processed by an affected TensorFlow Lite GatherNd implementation. The cited impact is denial of service through a crash condition.
Researcher notes
Evidence is limited to the TensorFlow advisory, CVE record, and fixing commit. The sources name the root cause, affected version ranges, and target fixed releases, but do not provide evidence of exploitation in the wild or broader product impact beyond TensorFlow/TFLite.
Mitigation direction
- Inventory TensorFlow and TFLite versions across applications, containers, mobile builds, and edge devices.
- Upgrade to TensorFlow 2.5.0 or patched supported releases 2.4.2, 2.3.3, 2.2.3, or 2.1.4.
- Restrict loading of untrusted TFLite models until affected runtimes are patched.
- Use vendor guidance if maintaining older unsupported TensorFlow branches.
- Prefer signed, curated model artifacts in production ML pipelines.
Validation and detection
- Confirm no deployed TensorFlow versions match the listed affected ranges.
- Check dependency lockfiles, container images, and mobile build artifacts for bundled TensorFlow Lite.
- Review whether any service accepts user-supplied or partner-supplied TFLite models.
- Run regression tests around model loading and GatherNd workloads after upgrading.
- Document compensating controls where immediate upgrade is not possible.
Public sources used
Generated from the cited source records. This long-tail analysis has not been individually reviewed by a named human.
Conservative CVE-to-ATT&CK context
These mappings and lookup hints may be relevant to the vulnerability behavior, CWE, affected product, or exposure path. Glexia-inferred context is not an official MITRE, ATT&CK, CWE, or CVE Program mapping.
ATT&CK lookup starting points
Use these exact CWE pages and searches to review the Glexia ATT&CK library from this CVE's weakness and description context.
CWE-369: Exact CWE lookup
Use the exact CWE identifier as the starting point before reviewing related ATT&CK behavior. Open the exact CWE lookup page first, then review the ATT&CK searches from that MITRE weakness context. This is a Glexia lookup hint, not an official ATT&CK mapping.
Open ATT&CK lookupCVE-2021-29589 mapping review
Open the CVE-to-ATT&CK bridge for reviewed, inferred, or future official mappings tied to this CVE.
Open ATT&CK lookup- Severity
- Low
- CVSS
- 2.5 (3.1)
- Known Exploited
- No
- Published
Vector: CVSS:3.1/AV:L/AC:H/PR:L/UI:N/S:U/C:N/I:N/A:L
CNA and ADP enrichment extracted from CVE v5
These fields come from the CVE record and ADP containers, not from Glexia's Take. They preserve time-varying source decisions such as CISA SSVC, KEV status, CVSS metrics, and provider references.
CVSS vector scores
1 official scoreWe collect every scored CVSS vector available in the official CNA and ADP containers. When more than one version is present, the table keeps the source vectors side by side instead of collapsing them into the highest score.
CVSS:3.1/AV:L/AC:H/PR:L/UI:N/S:U/C:N/I:N/A:L11.4Primary CVE scoreVulnerability scoring details
Base CVSS 3.1 score
2.5LowVector: CVSS:3.1/AV:L/AC:H/PR:L/UI:N/S:U/C:N/I:N/A:L
Source materials
- CVE List V5 sourceCVE List V5
- https://github.com/tensorflow/tensorflow/security/advisories/GHSA-3w67-q784-6w7cCVE reference · x_refsource_CONFIRM
- https://github.com/tensorflow/tensorflow/commit/8e45822aa0b9f5df4b4c64f221e64dc930a70a9dCVE reference · x_refsource_MISC
Products and packages named in the record
CWE details
CWE links open Glexia weakness intelligence pages with official CWE context, developer remediation guidance, and related CVE mappings.
Divide By Zero
Divide By Zero represents a recurring weakness pattern that can create exploitable paths when design, validation, or implementation controls are missing.