Issabel PBX Hard-Coded JWT Signing Key Remote Code Execution
First seen Sep 17, 2026 · Updated Sep 17, 2026 · CVSS 9.8
The Issabel Framework, used by Issabel PBX software, ships with a hard-coded HS256 JWT signing key in pbxapi index.php that is identical across all installations, allowing unauthenticated attackers to forge valid API tokens. Using forged tokens, attackers can invoke the manager originate endpoint with the System application parameter to execute arbitrary OS commands as the Asterisk user, and Shadowserver has already observed active exploitation in the wild as of September 2026.
Technical Analysis
The vulnerability stems from a static, publicly-known HS256 secret embedded in pbxapi's index.php prior to commit b97dbaf, which is used to sign and validate JWT bearer tokens for the PBX API. Because the key is identical on every deployment, an attacker with no credentials can craft a valid signed token and authenticate as a privileged API user. With a forged token, the attacker can call the Asterisk manager originate endpoint and pass the System application parameter, which Asterisk executes as an OS-level shell command under the asterisk service account, yielding remote code execution with CVSS 9.8. This is a classic cryptographic key-management failure (hard-coded secret) combined with unsafe exposure of a command-execution primitive (AMI originate/System) via a web API. While Issabel PBX itself is not an AI agent framework, organizations that deploy voice/IVR agents, call-routing bots, or AI-driven contact-center automation on top of Issabel-based Asterisk infrastructure could see their agent hosts compromised, credentials and API keys on the box exfiltrated, and downstream agent tool-use pipelines hijacked via lateral movement from the compromised PBX host.
Affected Systems
Issabel Framework / Issabel PBX web framework (pbxapi component) prior to commit b97dbaf; any Issabel PBX installation exposing the pbxapi/index.php endpoint to the network, including default and unpatched builds using Asterisk as the backend telephony engine.
Indicators of Compromise
- Exploitation reported by Shadowserver Foundation starting 2026-09-09
- Anomalous requests to pbxapi/index.php with crafted Authorization: Bearer JWT headers
- Manager API 'originate' calls invoking the 'System' application parameter
- Unexpected shell command execution under the 'asterisk' user process tree
- (Specific hashes/IPs/domains not disclosed in source data; consult Shadowserver feeds for current IOC lists)
Remediation Steps
- 1
Apply the patch / upgrade
Update Issabel Framework to a version containing commit b97dbaf or later, which removes the hard-coded JWT signing key and enforces per-installation secret generation.
- 2
Rotate JWT signing secrets
Manually regenerate and randomize the JWT signing key on all existing Issabel deployments, even after patching, to invalidate any previously forged or leaked tokens.
- 3
Restrict network exposure
Ensure pbxapi and the Asterisk Manager Interface (AMI) are not exposed to the public internet; restrict access via firewall rules, VPN, or network segmentation to trusted management hosts only.
- 4
Disable or restrict dangerous AMI applications
Limit or disable the ability to invoke the 'System' application via the manager originate endpoint, or implement strict allow-lists for permitted AMI actions.
- 5
Monitor and hunt for compromise
Review Asterisk and web server logs for anomalous API authentication and originate calls, and check for signs of unauthorized OS command execution as the asterisk user, especially since exploitation began before 2026-09-09.
- 6
Audit downstream integrations
If AI voice agents, IVR bots, or automation pipelines integrate with the affected PBX, rotate any API keys or credentials stored on or accessible from the host and audit for lateral movement into agent infrastructure.
CVE / Advisory IDs
Industries Most Exposed
Respond to this threat
Pro subscribers get a full AI-generated incident-response playbook for this threat — detection, containment, eradication, and recovery steps — plus an unlimited AI Threat Advisor for questions about your environment.