Executive assessment
Today's brief leads with Node2Shell (Brotli): now in CISA KEV; NHS England cyber alert CC-4881. All 15 selected findings retain their own technical scope, action, observed status, and evidence limits.
Also today: 2 more AIOHTTP CVEs (CVE-2026-34513, CVE-2026-34515) in the same disclosure wave as the AIOHTTP card of 2026-10-20; none reported exploited; carried as a note rather than a finding.
Finding 01 — Node2Shell (Brotli): now in CISA KEV; NHS England cyber alert CC-4881
What changed: Follow-up to the 2026-10-24 card. NHS England published cyber alert CC-4881. Out-of-bounds write in .NET allows an unauthorized attacker to elevate privileges over a network.
Technical evidence: CVE-2026-111497; CVSS v4.0 9.2; weakness CWE-787, CWE-252; technical confidence High.
Why it matters: Reported active exploitation elevates this above routine patching: validate exposure immediately, remediate, and assess for prior compromise.
Observed status: Observed in-the-wild exploitation is confirmed.
Analyst note: For PAN-OS 11.2.4 or later, disable Brotli inspection globally with set system setting brotli-encoding disable. ([security.paloaltonetworks.com](https://security.paloaltonetworks.com/CVE-2026-111497), [cveawg.mitre.org](https://cveawg.mitre.org/api/cve/CVE-2026-111497))
Affected: Google Brotli versions 1.0.3 through 1.2.0 on all platforms. ([security.paloaltonetworks.com](https://security.paloaltonetworks.com/CVE-2026-111497))
Fix: Fixed upstream releases are Brotli 1.0.10, 1.1.1 or 1.2.1. ([security.paloaltonetworks.com](https://security.paloaltonetworks.com/CVE-2026-111497))
Panel assessment: Patch Google Brotli now wherever it decodes untrusted compressed streams, because exploitation is already happening in the wild and Brotli is an embedded dependency rather than a single product, so the affected range of 1.0.3 through 1.2.0 reaches web servers, runtimes, client software and security appliances that inspect traffic. The practical path is a crafted compressed response or request that triggers the out of bounds write inside the decoding process, giving the attacker execution in whatever context that process runs, which on a proxy or inspection engine means the credentials, session material and decrypted traffic that process already holds. (priority: patch now)
Action: Map CVE-2026-111497 to owned assets, confirm the affected and fixed versions in the cited source, deploy the applicable remediation, and retain evidence of the exposure decision.
[Evidence source: msrc.microsoft.com](https://msrc.microsoft.com/update-guide/vulnerability/CVE-2026-111497)
Finding 02 — Multiple vulnerabilities in Mitsubishi Electric GENESIS, GENESIS64, ICONICS Suite and seven more products
What changed: Successful exploitation of these vulnerabilities could allow an attacker to execute arbitrary code (CVE-2023-4863), bypass authentication (CVE-2026-16436 and CVE-2026-16437), or perform arbitrary operations (CVE-2026-16438), resulting in information disclosure, tampering with, destruction, or deletion of information, and a denial-of-service (DoS) condition of the system. CVE coverage: CVE-2023-4863, CVE-2026-16436, CVE-2026-16437, CVE-2026-16438.
Technical evidence: CVE-2026-16436; CVSS v4.0 9.2; weakness CWE-640; technical confidence High.
Why it matters: The Critical priority moves this beyond routine tracking and requires exposure validation, remediation, and compromise assessment.
Analyst note: Block remote login from untrusted networks and hosts and from non-administrator users; when internet access is required, use a firewall or VPN and allow remote login only to an administrator. ([cisa.gov](https://www.cisa.gov/news-events/ics-advisories/icsa-26-300-02), [cisa.gov](https://www.cisa.gov/known-exploited-vulnerabilities-catalog), [cisa.gov](https://www.cisa.gov/sites/default/files/feeds/known_exploited_vulnerabilities.json))
Affected: CVE-2023-4863: GENESIS64 and ICONICS Suite 10.97 to 10.97.2. CVE-2026-16436 and CVE-2026-16437: GENESIS 11.03 and prior. ([cisa.gov](https://www.cisa.gov/news-events/ics-advisories/icsa-26-300-02))
Fix: CVE-2023-4863: GENESIS64 and ICONICS Suite 10.97.3 or later. CVE-2026-16436 and CVE-2026-16437: GENESIS64, ICONICS Suite, Hyper Historian, AnalytiX and MobileHMI 10.98 or later. ([cisa.gov](https://www.cisa.gov/news-events/ics-advisories/icsa-26-300-02))
Panel assessment: Mitsubishi Electric GENESIS, GENESIS64 and the wider ICONICS family need immediate attention because one of the flaws is already exploited in the wild and the rest break authentication and privilege boundaries across supervisory HMI, historian and analytics components that sit close to process control. The plausible chain is remote login or an exposed operator workstation, abuse of the weak challenge-response recovery to land as a valid user, then arbitrary operations against historian records, dashboards and configuration, giving tampering, deletion or loss of operator visibility. (priority: patch now)
Action: Map CVE-2026-16436 to owned assets, confirm the affected and fixed versions in the cited source, deploy the applicable remediation, and retain evidence of the exposure decision.
[Evidence source: cisa.gov](https://www.cisa.gov/news-events/ics-advisories/icsa-26-300-02)
Finding 03 — Multiple vulnerabilities in EPICS Base
What changed: Successful exploitation of these vulnerabilities could allow an attacker to cause a denial-of-service condition, corrupt memory, disclose sensitive information, or potentially execute arbitrary code. CVE coverage: CVE-2026-114595, CVE-2026-114596, CVE-2026-114597, CVE-2026-114598.
Technical evidence: CVE-2026-114595; CVSS v3.1 9.8; weakness CWE-190; technical confidence High.
Why it matters: The Critical priority moves this beyond routine tracking and requires exposure validation, remediation, and compromise assessment.
Analyst note: Keep control-system devices and systems off the public internet. Place control-system networks and remote devices behind firewalls and isolate them from business networks. CISA reports that no known public exploitation specifically targeting these vulnerabilities has been reported. ([cisa.gov](https://www.cisa.gov/news-events/ics-advisories/icsa-26-300-01))
Affected: EPICS Base releases up to and including 7.0.10. ([cisa.gov](https://www.cisa.gov/news-events/ics-advisories/icsa-26-300-01))
Fix: EPICS Base 7.0.11. ([cisa.gov](https://www.cisa.gov/news-events/ics-advisories/icsa-26-300-01))
Action: Map CVE-2026-114595 to owned assets, confirm the affected and fixed versions in the cited source, deploy the applicable remediation, and retain evidence of the exposure decision.
[Evidence source: cisa.gov](https://www.cisa.gov/news-events/ics-advisories/icsa-26-300-01)
Finding 04 — vLLM: Request-controlled mm_processor_kwargs.code_revision allows remote code execution
What changed: GitHub Advisory Database published GHSA-h3rc-6mm3-gc2m for CVE-2026-114053 (vLLM): Code Injection, CVSS v4.0 9.2; affected: vLLM versions 0.7.3 and later, before 0.31.0.
Technical evidence: CVE-2026-114053; CVSS v4.0 9.2; weakness CWE-94; technical confidence High.
Why it matters: The Critical priority moves this beyond routine tracking and requires exposure validation, remediation, and compromise assessment.
Analyst note: Disable trust_remote_code; if it must remain enabled, require authentication for /tokenize at an external proxy. ([osv.dev](https://osv.dev/vulnerability/GHSA-h3rc-6mm3-gc2m), [cveawg.mitre.org](https://cveawg.mitre.org/api/cve/CVE-2026-114053))
Affected: vLLM versions 0.7.3 and later, before 0.31.0. ([cveawg.mitre.org](https://cveawg.mitre.org/api/cve/CVE-2026-114053))
Fix: vLLM 0.31.0. ([cveawg.mitre.org](https://cveawg.mitre.org/api/cve/CVE-2026-114053))
Action: Map CVE-2026-114053 to owned assets, confirm the affected and fixed versions in the cited source, deploy the applicable remediation, and retain evidence of the exposure decision.
[Evidence source: github.com](https://github.com/advisories/GHSA-h3rc-6mm3-gc2m)
Finding 05 — LiteLLM: Privilege Escalation to Proxy Admin via Cross-Domain Reuse of the Salt Key
What changed: GitHub Advisory Database published GHSA-7hp6-4w63-5g45 for CVE-2026-114612 (LiteLLM): Improper Privilege Management, CVSS v3.1 9.9; affected: Versions 1.87.0 through 1.90.x are exploitable only if EXPERIMENTAL_UI_LOGIN=true was explicitly set.
Technical evidence: CVE-2026-114612; CVSS v3.1 9.9; weakness CWE-269, CWE-345, CWE-441; technical confidence High.
Why it matters: The Critical priority moves this beyond routine tracking and requires exposure validation, remediation, and compromise assessment.
Observed status: Observed in-the-wild exploitation status is unknown.
Analyst note: If patching is not immediately possible, set EXPERIMENTAL_UI_LOGIN=false; this disables the vulnerable authentication path but breaks CLI SSO and Claude Code gateway login. ([github.com](https://github.com/BerriAI/litellm/security/advisories/GHSA-7hp6-4w63-5g45))
Affected: Versions 1.87.0 through 1.90.x are exploitable only if EXPERIMENTAL_UI_LOGIN=true was explicitly set ([github.com](https://github.com/BerriAI/litellm/security/advisories/GHSA-7hp6-4w63-5g45))
Fix: 1.100.4, 1.101.3, 1.102.2, 1.103.1, and 1.104.0rc2 ([github.com](https://github.com/BerriAI/litellm/security/advisories/GHSA-7hp6-4w63-5g45))
Action: Map CVE-2026-114612 to owned assets, confirm the affected and fixed versions in the cited source, deploy the applicable remediation, and retain evidence of the exposure decision.
[Evidence source: github.com](https://github.com/advisories/GHSA-7hp6-4w63-5g45)
Finding 06 — Multiple vulnerabilities in Authlib
What changed: Authlib carries 8 CVEs across 3 advisories: Improper Authentication; Incomplete List of Disallowed Inputs; ID token can disable its own nonce validation via the nonce_supported claim; Default JWT bearer grant accepts assertions for the wrong audience; OpenID implicit grant checks raw scope instead of allowed scope before issuing ID tokens; is_secure_transport loopback exemption can be bypassed with a userinfo component, allowing cleartext-HTTP URLs to non-loopback hosts; empty signatures validation in JWS JSON deserialization. CVE coverage: CVE-2026-114620, CVE-2026-114617, CVE-2026-114618, CVE-2026-114619, CVE-2026-114614, CVE-2026-96760, CVE-2026-114616, CVE-2026-114615.
Technical evidence: CVE-2026-114620; CVSS v3.1 5.4; weakness CWE-918; technical confidence High.
Why it matters: The Critical priority moves this beyond routine tracking and requires exposure validation, remediation, and compromise assessment.
Observed status: Observed in-the-wild exploitation status is unknown.
Action: Map CVE-2026-114620 to owned assets, confirm the affected and fixed versions in the cited source, deploy the applicable remediation, and retain evidence of the exposure decision.
Evidence limits: fixed version or patch state unknown
[Evidence source: github.com](https://github.com/advisories/GHSA-x4rv-c3wx-4cr8)
Finding 07 — better-auth: Improper Authentication
What changed: GitHub Advisory Database published GHSA-965c-763c-88jm for CVE-2026-114607 (better-auth): Improper Authentication, CVSS v3.1 9.1; affected: better-auth 1.4.0-beta.18 through versions before 1.7.7, when the Magic Link plugin and social or Generic OAuth sign-in are used.
Technical evidence: CVE-2026-114607; CVSS v3.1 9.1; weakness CWE-287; technical confidence High.
Why it matters: The Critical priority moves this beyond routine tracking and requires exposure validation, remediation, and compromise assessment.
Observed status: Observed in-the-wild exploitation status is unknown.
Analyst note: Set account.storeStateStrategy to "cookie", or disable Magic Link sign-in or social and Generic OAuth sign-in. If OAuth Proxy is used, first configure a dedicated OAuth Proxy secret across participating environments that differs from each environment's global Better Auth secret. ([github.com](https://github.com/better-auth/better-auth/security/advisories/GHSA-965c-763c-88jm), [api.osv.dev](https://api.osv.dev/v1/vulns/CVE-2026-114607))
Affected: better-auth 1.4.0-beta.18 through versions before 1.7.7, when the Magic Link plugin and social or Generic OAuth sign-in are used. ([github.com](https://github.com/better-auth/better-auth/security/advisories/GHSA-965c-763c-88jm))
Fix: better-auth 1.7.7 ([api.osv.dev](https://api.osv.dev/v1/vulns/CVE-2026-114607))
Action: Map CVE-2026-114607 to owned assets, confirm the affected and fixed versions in the cited source, deploy the applicable remediation, and retain evidence of the exposure decision.
[Evidence source: github.com](https://github.com/advisories/GHSA-965c-763c-88jm)
Finding 08 — Multiple vulnerabilities in MLflow CrewAI
What changed: MLflow CrewAI carries 3 CVEs across 3 advisories: Code Injection; Path Traversal; Exposure of Sensitive Information to an Unauthorized Actor. CVE coverage: CVE-2026-116692, CVE-2026-116690, CVE-2026-116691.
Technical evidence: CVE-2026-116692; CVSS v3.1 9.8; weakness CWE-94, CWE-470; technical confidence High.
Why it matters: The Critical priority moves this beyond routine tracking and requires exposure validation, remediation, and compromise assessment.
Observed status: Observed in-the-wild exploitation status is unknown.
Action: Map CVE-2026-116692 to owned assets, confirm the affected and fixed versions in the cited source, deploy the applicable remediation, and retain evidence of the exposure decision.
Evidence limits: fixed version or patch state unknown
[Evidence source: github.com](https://github.com/advisories/GHSA-26p8-2jq9-3vq9)
Finding 09 — Multiple vulnerabilities in Apple iOS
What changed: Apple iOS carries 2 CVEs across 2 advisories: Out-of-bounds Write. CVE coverage: CVE-2026-86950, CVE-2026-20649.
Technical evidence: CVE-2026-86950; CVSS v3.1 8.8; weakness CWE-787; technical confidence High.
Why it matters: Reported active exploitation elevates this above routine patching: validate exposure immediately, remediate, and assess for prior compromise.
Observed status: Observed in-the-wild exploitation is confirmed.
Action: Map CVE-2026-86950 to owned assets, confirm the affected and fixed versions in the cited source, deploy the applicable remediation, and retain evidence of the exposure decision.
Evidence limits: fixed version or patch state unknown
[Evidence source: bleepingcomputer.com](https://www.bleepingcomputer.com/news/apple/apple-backports-coregraphics-zero-day-patches-to-older-iphones/)
Finding 10 — Multiple vulnerabilities in Microsoft Edge
What changed: CVE coverage: CVE-2026-62765, CVE-2026-20942, CVE-2026-88078, CVE-2026-96196. msrc.microsoft.com records CVE-2026-62765 (Microsoft Edge): Type Confusion, CVSS v3.1 7.5.
Technical evidence: CVE-2026-62765; CVSS v3.1 7.5; weakness CWE-843; technical confidence Medium.
Why it matters: The High priority moves this beyond routine tracking and requires exposure validation, remediation, and compromise assessment.
Analyst note: Until patching, use Edge URLBlocklist to block identified malicious sites. Set ExtensionInstallBlocklist to '' and allow-list only required signed extensions. Microsoft reports all four vulnerabilities as not exploited at original publication. ([msrc.microsoft.com](https://msrc.microsoft.com/update-guide/vulnerability/CVE-2026-62765), [msrc.microsoft.com](https://msrc.microsoft.com/update-guide/vulnerability/CVE-2026-20942), [msrc.microsoft.com](https://msrc.microsoft.com/update-guide/vulnerability/CVE-2026-88078), +3 more)
Fix: CVE-2026-62765: build 149.0.4022.62; CVE-2026-20942, CVE-2026-88078 and CVE-2026-96196: build 156.0.4314.42. ([msrc.microsoft.com](https://msrc.microsoft.com/update-guide/vulnerability/CVE-2026-62765), [msrc.microsoft.com](https://msrc.microsoft.com/update-guide/vulnerability/CVE-2026-20942), [msrc.microsoft.com](https://msrc.microsoft.com/update-guide/vulnerability/CVE-2026-88078), +1 more)
Panel assessment: Microsoft Edge sits on nearly every Windows endpoint and renders untrusted content by default, so the type confusion flaw gives a drive-by route to code execution in the browser context even though Microsoft reports none of the four as exploited yet. A plausible chain starts with a crafted page and a popup click, where the spoofing and feature bypass issues make the lure more convincing and weaken the warnings a user would otherwise see, putting session tokens, saved credentials and access to internal web applications at risk. Check first which build each estate is actually running, because the fixes land in two different builds and a fleet patched only to the newer train may still carry the remote code execution gap. (priority: this week)
Action: Map CVE-2026-62765 to owned assets, confirm the affected and fixed versions in the cited source, deploy the applicable remediation, and retain evidence of the exposure decision.
[Evidence source: msrc.microsoft.com](https://msrc.microsoft.com/update-guide/vulnerability/CVE-2026-62765)
Finding 11 — Langflow: Authorization Bypass Through User-Controlled Key
What changed: Langflow is a tool for building and deploying AI-powered agents and workflows. Prior to version 1.5.1, the _read_flow helper in src/backend/base/langflow/api/v1/flows.py branched on the AUTO_LOGIN setting to decide whether to filter by user_id.
Technical evidence: CVE-2026-34046; CVSS v4.0 8.7; weakness CWE-639, CWE-862; technical confidence Medium.
Why it matters: The High priority moves this beyond routine tracking and requires exposure validation, remediation, and compromise assessment.
Analyst note: If patching is not immediately possible, temporarily block GET, PATCH and DELETE requests to /api/v1/flow/{flow_id} at the ingress. CISA reports exploitation as none. ([github.com](https://github.com/langflow-ai/langflow/security/advisories/GHSA-8c4j-f57c-35cf), [services.nvd.nist.gov](https://services.nvd.nist.gov/rest/json/cves/2.0?cveId=CVE-2026-34046), [api.osv.dev](https://api.osv.dev/v1/vulns/GHSA-8c4j-f57c-35cf))
Affected: PyPI package langflow-base versions earlier than 0.5.1 are affected. ([api.osv.dev](https://api.osv.dev/v1/vulns/GHSA-8c4j-f57c-35cf))
Fix: PyPI package langflow-base version 0.5.1 contains the fix. ([api.osv.dev](https://api.osv.dev/v1/vulns/GHSA-8c4j-f57c-35cf))
Action: Map CVE-2026-34046 to owned assets, confirm the affected and fixed versions in the cited source, deploy the applicable remediation, and retain evidence of the exposure decision.
[Evidence source: nvd.nist.gov](https://nvd.nist.gov/vuln/detail/CVE-2026-34046)
Finding 12 — Hi.Events: SQL Injection
What changed: Hi.Events is an open-source event management and ticket selling platform. From version 0.8.0-beta.1 to before version 1.7.1-beta, multiple repository classes pass the user-supplied sort_by query parameter directly to Eloquent's orderBy() without validation, enabling SQL injection.
Technical evidence: CVE-2026-34455; CVSS v4.0 8.7; weakness CWE-89; technical confidence Medium.
Why it matters: The High priority moves this beyond routine tracking and requires exposure validation, remediation, and compromise assessment.
Analyst note: Enforce an allowlist of accepted sort_by values at the WAF or reverse proxy, and block values containing stacked-query delimiters or SQL comments. Alert on sort_by requests containing pg_sleep, semicolons or --. The Hi.Events GitHub Security Advisory publishes an authenticated-organiser time-based blind SQL injection proof of concept. ([github.com](https://github.com/HiEventsDev/Hi.Events/security/advisories/GHSA-2qcp-24fh-fx6p))
Panel assessment: Hi.Events deployments should be patched to 1.7.1-beta now, because a working proof of concept is already public and the organiser role it needs is a role many event teams hand out widely, so the real barrier is lower than authentication implies. The practical path is a crafted sort_by value on a listing endpoint such as attendees, where PostgreSQL stacked queries let a request move beyond reading rows into writing them, putting attendee and ticketing data and whatever else the application database account can touch at risk. First confirm the running version against the patched commit, then search proxy and application logs for historical sort_by values containing semicolons, SQL comment markers or pg_sleep, since the time based technique leaves slow requests rather than errors. (priority: patch now)
Action: Map CVE-2026-34455 to owned assets, confirm the affected and fixed versions in the cited source, deploy the applicable remediation, and retain evidence of the exposure decision.
[Evidence source: nvd.nist.gov](https://nvd.nist.gov/vuln/detail/CVE-2026-34455)
Finding 13 — Atlantis: OS command injection through PR-controlled Terraform workspace enables RCE during plan
What changed: GitHub Advisory Database published GHSA-rjh2-4xgw-p2hj for CVE-2026-116676 (Atlantis): OS Command Injection, CVSS v3.1 8.8; affected: >= 0.4.0, <= 0.47.0.
Technical evidence: CVE-2026-116676; CVSS v3.1 8.8; weakness CWE-78; technical confidence High.
Why it matters: The High priority moves this beyond routine tracking and requires exposure validation, remediation, and compromise assessment.
Analyst note: Enforce strict workspace-name validation that rejects shell metacharacters. Disable --allow-fork-prs; because same-repository branches do not require this flag, also prevent untrusted contributors from triggering Atlantis plans. ([github.com](https://github.com/runatlantis/atlantis/security/advisories/GHSA-rjh2-4xgw-p2hj))
Affected: >= 0.4.0, <= 0.47.0 ([github.com](https://github.com/runatlantis/atlantis/security/advisories/GHSA-rjh2-4xgw-p2hj))
Fix: 0.47.1 ([github.com](https://github.com/runatlantis/atlantis/security/advisories/GHSA-rjh2-4xgw-p2hj))
Action: Map CVE-2026-116676 to owned assets, confirm the affected and fixed versions in the cited source, deploy the applicable remediation, and retain evidence of the exposure decision.
[Evidence source: github.com](https://github.com/advisories/GHSA-rjh2-4xgw-p2hj)
Finding 14 — CTFer.io's CTFd Terraform Provider: Insufficiently Protected Credentials
What changed: GitHub Advisory Database published GHSA-vg84-c5mg-76xh for CVE-2026-116683 (CTFer.io): Insufficiently Protected Credentials, CVSS v4.0 7.1; affected: v0.1.5 through v2.8.1, inclusive.
Technical evidence: CVE-2026-116683; CVSS v4.0 7.1; weakness CWE-522; technical confidence Medium.
Why it matters: The High priority moves this beyond routine tracking and requires exposure validation, remediation, and compromise assessment.
Observed status: Observed in-the-wild exploitation status is unknown.
Analyst note: Keep Terraform state in safe environments and limit access to a small set of users. ([github.com](https://github.com/ctfer-io/terraform-provider-ctfd/security/advisories/GHSA-vg84-c5mg-76xh))
Affected: v0.1.5 through v2.8.1, inclusive. ([github.com](https://github.com/ctfer-io/terraform-provider-ctfd/security/advisories/GHSA-vg84-c5mg-76xh))
Fix: v2.8.2. ([github.com](https://github.com/ctfer-io/terraform-provider-ctfd/security/advisories/GHSA-vg84-c5mg-76xh))
Panel assessment: Upgrading the CTFer.io CTFd Terraform Provider is worth scheduling rather than rushing, because the fix in v2.8.2 stops new leakage but does nothing about passwords already written in clear into existing state, so remediation is really a credential rotation job. The practical path needs no exploit, only read access to state wherever it lives, which in most pipelines means remote backends, CI artefacts and any copy committed to version control, and what it reaches is team credentials that may be reused elsewhere. If the provider is in use, first find every historical state copy produced by v0.1.5 through v2.8.1, including backend version history and pipeline logs, and treat the passwords in them as already disclosed. (panel view at moderate confidence; priority: this week)
Action: Map CVE-2026-116683 to owned assets, confirm the affected and fixed versions in the cited source, deploy the applicable remediation, and retain evidence of the exposure decision.
[Evidence source: github.com](https://github.com/advisories/GHSA-vg84-c5mg-76xh)
Finding 15 — MongoDB Python Driver: Out-of-bounds read in BSON decoding via a malformed Regex element
What changed: GitHub Advisory Database published GHSA-98qm-wvgf-w4cf for CVE-2026-106435 (MongoDB Python Driver): Out-of-bounds Read, CVSS v4.0 5.9; affected: All releases through 4.18.2 are affected when the C extension is loaded.
Technical evidence: CVE-2026-106435; CVSS v4.0 5.9; weakness CWE-125; technical confidence Medium.
Why it matters: The Medium priority requires exposure validation and scheduled remediation through normal change control.
Observed status: Observed in-the-wild exploitation status is unknown.
Analyst note: Decode untrusted BSON only from NUL-terminated buffers, such as bytes or bytearray, and copy mmap input first, for example bytes(mm[:size]). ([github.com](https://github.com/advisories/GHSA-98qm-wvgf-w4cf))
Affected: All releases through 4.18.2 are affected when the C extension is loaded. ([github.com](https://github.com/advisories/GHSA-98qm-wvgf-w4cf))
Fix: PyMongo 4.18.3. ([github.com](https://github.com/advisories/GHSA-98qm-wvgf-w4cf))
Action:** Map CVE-2026-106435 to owned assets, confirm the affected and fixed versions in the cited source, deploy the applicable remediation, and retain evidence of the exposure decision.
[Evidence source: github.com](https://github.com/advisories/GHSA-98qm-wvgf-w4cf)