Skip to main content

Katanemo Plano EUVDEUVD-2026-97014

| CVE-2026-108863 HIGH
Missing Authentication for Critical Function (CWE-306)
2026-10-11 VulnCheck GHSA-54m9-qc5w-93rc
8.7
CVSS 4.0 · Vendor: VulnCheck
View

Attack ChainAIDerived

Hypothetical attack flow derived from CVE metadata – not a confirmed exploit chain

  1. Access
    Scan for exposed Envoy admin port 9901
  2. Delivery
    Send unauthenticated GET /config_dump
  3. Exploit
    Parse returned WASM filter configuration
  4. Execution
    Extract plaintext LLM provider API keys
  5. Impact
    Abuse stolen keys against AI provider accounts

Vulnerability Assessment

Weakness
CWE
OWASP Top 10:2025
A07:2025 - Authentication Failures
ASVS 5.0
V6 - Authentication Authentication required for critical functions
MITRE ATT&CK

1 technique across 1 tactic: Initial Access

The technique list with ATT&CK links and the kill-chain view is part of the Basic plan and up.

Exploitation Conditions

Exploitation requires the Envoy admin interface to be running and network-reachable on TCP port 9901 and at least one LLM provider API key to be configured in the WASM filter. Per the referenced config/envoy.template.yaml in Plano 0.4.37, the admin interface is bound to all host interfaces (0.0.0.0) by default, so stock deployments are exposed without any special configuration - there are no authentication, user-interaction, or non-default-setting prerequisites. The primary limiting factor is network topology: if operators firewall, bind to loopback, or otherwise isolate port 9901 from untrusted networks, remote exploitation is prevented. The flaw also only yields value when real provider keys are present in the WASM filter configuration.

Impact

The CVSS 4.0 vector (AV:N/AC:L/AT:N/PR:N/UI:N) describes a textbook low-effort, high-reward exposure: an unauthenticated network attacker needs only to reach TCP/9901 and issue a single GET to /config_dump, with no user interaction, no authentication, and no special attack requirements. The impact profile is narrow but severe on one axis - VC:H with VI:N and VA:N - meaning this is purely a confidentiality (secret-disclosure) flaw, not a code-execution or integrity/availability issue, which is why the base score lands at 8.7 rather than near 10. No EPSS score, CISA KEV listing, or public POC is present in the provided intelligence, so there is no public exploit identified at time of analysis and no evidence of active exploitation; the finding originates from VulnCheck research rather than observed attacks. The signals are internally consistent - the network vector, low complexity, and no-privilege requirement all align with the description of an admin port bound to all interfaces - so there is no contradiction to flag between CVSS and the narrative. The practical priority is high for anyone running Plano with 9901 exposed to untrusted networks because exploitation is trivial and yields live LLM provider API keys (enabling downstream billing abuse and lateral access to AI accounts), but the real-world blast radius is tempered by the product's niche footprint and by whether operators have network-isolated the admin port. Absent EPSS/KEV data, treat widespread opportunistic exploitation as plausible-but-unproven rather than confirmed.

Exploit ScenarioDerived

An attacker scanning for exposed Envoy admin interfaces finds a Plano instance with port 9901 reachable and sends an unauthenticated HTTP request to /config_dump. The returned configuration includes the WASM filter settings, from which the attacker reads the configured LLM provider API keys in plaintext and reuses them to bill against or access the victim's AI accounts. No public exploit code is required - the attack is a single standard HTTP GET against a default-exposed endpoint.

This Week Schedule remediation this week. Why: HIGH severity (CVSS 7.0-8.9).
No patch listed yet 0.4% exploitation probability (EPSS)

Get an e-mail when this CVE is exploited, patched or re-scored. Free account.

Alert me

AnalysisAIv1 – Oct 11, 2026

Unauthenticated information disclosure in Katanemo Plano through 0.4.37 allows any network attacker who can reach TCP port 9901 to read plaintext LLM provider API keys by issuing a single GET request to the Envoy admin interface's /config_dump endpoint. The Envoy admin listener is bound to all host interfaces (0.0.0.0) in the shipped config/envoy.template.yaml, so stock deployments are exposed by default; the only real barriers are network reachability to port 9901 and the presence of at least one configured provider key, and the CVSS 4.0 vector (AV:N/AC:L/AT:N/PR:N/UI:N, C:H) reflects a low-effort, high-reward credential theft with no authentication or user interaction required. No public exploit code has been identified at time of analysis and the issue is not listed in CISA KEV, although the flaw has been publicly documented by VulnCheck.

RemediationAI

No vendor-released patch was identified at time of analysis, so remediation depends on compensating controls applied at the deployment layer. The most direct fix is to edit config/envoy.template.yaml (or the generated Envoy config) to bind the admin interface to loopback only - for example an admin address of 127.0.0.1:9901 - which removes remote reachability; the trade-off is that operators lose remote access to admin diagnostics and must tunnel or exec into the host to use them. Alternatively, block TCP 9901 at the host firewall, security group, or network policy so only trusted management hosts can reach it, accepting that any host with permitted access retains full read of the config; and where remote admin access is needed, front it with an authenticating reverse proxy and restrict /config_dump specifically. Because the interface leaks secrets with no authentication, treat every API key configured in the Plano WASM filter as compromised on any deployment that was exposed and rotate the affected provider credentials immediately, then audit provider logs for unauthorized usage. Monitor project releases at https://github.com/katanemo/plano and the advisory at https://www.vulncheck.com/advisories/katanemo-plano-through-0.4.37-missing-authentication-on-envoy-admin-interface for a fixed version.

Affected VersionsENISA

Plano 0 ≤0.4.37

Affected Packages

euvd plano no fix

Source: OSV (Open Source Vulnerabilities)

Affects your stack

Affects 1 package in euvd

Checking this CVE against your own stacks and servers is part of the Basic plan and up.

Technical ContextAI · show details

Plano is Katanemo's Envoy-based proxy/gateway for LLM and agentic traffic, and like any Envoy deployment it ships an administrative interface - conventionally bound to port 9901 - that exposes diagnostic endpoints such as /config_dump, /clusters and /stats. In Plano 0.4.37 the admin block in config/envoy.template.yaml binds this interface to all host interfaces rather than to loopback, and the CLI config generator (cli/planoai/config_generator.py) populates the WASM filter configuration that carries the model provider credentials. Because Envoy's admin endpoints have no built-in authentication, /config_dump serializes the full running configuration - including the WASM filter's provider API keys - in plaintext to any requester. The root cause is CWE-306 (Missing Authentication for a Critical Function): a privileged management function is exposed on the network without any credential check, turning a routine operational port into a credential-disclosure surface for LLM provider secrets.

Know when Plano gets its next vulnerability

  • Watch products and get an email when something new is published
  • Check this CVE against your own stack and servers
  • Exploit and PoC references, threat intelligence from 22 sources

No credit card · 7-day full trial

Severity by source

Vendor (VulnCheck) PRIMARY 8.7 HIGH
CVSS:4.0/AV:N/AC:L/AT:N/PR:N/UI:N/VC:H/VI:N/VA:N/SC:N/SI:N/SA:N
vuln.today AI 7.5 HIGH

Unauthenticated network request to a default-exposed admin port (AV:N/AC:L/PR:N/UI:N) discloses plaintext secrets, so C:H with I:N/A:N and scope unchanged.

3.1 AV:N/AC:L/PR:N/UI:N/S:U/C:H/I:N/A:N
4.0 AV:N/AC:L/AT:N/PR:N/UI:N/VC:H/VI:N/VA:N/SC:N/SI:N/SA:N

Primary rating from Vendor (VulnCheck).

CVSS VectorVendor: VulnCheck

Attack Vector
Network
Attack Complexity
Low
Attack Requirements
None
Privileges Required
None
User Interaction
None
Vulnerable system: Confidentiality
High
Vulnerable system: Integrity
None
Vulnerable system: Availability
None
Subsequent system: Confidentiality
None
Subsequent system: Integrity
None
Subsequent system: Availability
None

DescriptionCVE.org

Katanemo Plano through 0.4.37 contains a missing authentication vulnerability that allows unauthenticated network attackers to access the Envoy admin interface, which is bound to all host interfaces on port 9901. Attackers can request the /config_dump endpoint to read configured LLM provider API keys in plaintext from the WASM filter configuration.

Executive ViewAI

Why It Matters

Katanemo Plano deployments up to and including version 0.4.37 ship with the Envoy admin interface bound to all host interfaces, allowing any network attacker who can reach TCP port 9901 to retrieve plaintext LLM provider API keys with a single unauthenticated request. Because no credentials or user interaction are required, every organization running a stock Plano deployment may have already exposed third-party AI provider keys to anything that can route to that port. Leadership should treat this as a credential-compromise event affecting AI service billing, model access, and vendor data-handling obligations rather than as a routine software defect.

Business Impact: High
Compensating Controls

No vendor-released patch identified at time of analysis. Interim mitigations: restrict TCP port 9901 to loopback or a trusted administrative network via host firewall, security group, or network policy; rebind or disable the Envoy admin listener in config/envoy.template.yaml; rotate all configured LLM provider API keys; segment Plano hosts from general network access; and monitor for GET requests to the Envoy /config_dump endpoint.

Threat Intelligence Overview

EPSS Score
0.4%
Top 65% exploitable
Prevalence
Low
niche open-source AI/LLM gateway, limited production footprint
Social Media
1
Sightings
Data Quality
5 sources
Confidence: Medium

How this decision was reached

The engine reads each signal below and applies a fixed list of rules from top to bottom. The first rule that matches sets the recommended action – like a statement, every line is shown.

SignalValue
Severity HIGH
CISA KEV No
Public POC No
EPSS 0.4%
Patch available No
Active threat No
Deciding rule
HIGH severity (CVSS 7.0-8.9)
What every action level means
Emergency Immediate action required. Criteria: CRITICAL + in CISA KEV (actively exploited) without patch, OR CRITICAL + EPSS > 30% without patch, OR CRITICAL + active threat signals without patch.
Act Now Prioritize within 24-48h. Criteria: in KEV with patch available, OR CRITICAL + public POC, OR HIGH + EPSS > 10%, OR CRITICAL + active threat, OR any CRITICAL severity.
This Week Schedule remediation this week. Criteria: HIGH severity (CVSS 7.0-8.9). Patch within normal change window.
This Month Plan within 30 days. Criteria: MEDIUM severity (CVSS 4.0-6.9), lower exploitation probability. Include in regular patch cycle.
Monitor Track and reassess. Criteria: LOW severity (CVSS 0.1-3.9), minimal exploitation risk. No immediate action needed.

SSVC AssessmentENISA

SSVC Decision· derived
Scheduled
Exploitation· derived
None
System Exposure· derived
Open
Automatable· derived
Yes
Human Impact
High

Priority Score

44
Low Medium High Critical
KEV: 0
EPSS: 0
CVSS: +44
PoC: 0
Weaponized: 0
Threat intel: 0
Gov advisory: 0

Attack Technique

Authentication bypass attacks exploit flaws in the verification mechanisms that control access to systems and applications.

This site uses cookies essential for authentication and security. No tracking or analytics cookies are used. Privacy Policy