/

CVE-2026-112380

Severity:

CRITICAL RISK

(9.9)

Missing Authorization (CWE-862) in Kiro Crew WhatsApp

CVE-2026-112380

Deferred

Vulnerability type

Category:

Missing Authorization

CWE:

CWE-862, CWE-863

Impact:

Remote Command Execution

Severity:

CRITICAL RISK

(9.9)

CVSS version:

3.1

Catch this CVE while code is being written

Detect vulnerable code directly in pull requests, so CVEs are fixed before merge.

Summary

Kiro Crew may fail to properly authorize messages received through an operator-configured WhatsApp group. A remote group member could potentially control other agent sessions and cause arbitrary commands to execute on the host running Kiro Crew.

Kiro Crew may fail to properly authorize messages received through an operator-configured WhatsApp group. A remote group member could potentially control other agent sessions and cause arbitrary commands to execute on the host running Kiro Crew.

Affected context

The issue affects deployments that use Kiro Crew's WhatsApp group channel to receive commands or control agent sessions. Risk is present when a user who is merely a member of an operator-configured group can send messages that are trusted across other agent sessions without adequate authorization or session isolation.

The issue affects deployments that use Kiro Crew's WhatsApp group channel to receive commands or control agent sessions. Risk is present when a user who is merely a member of an operator-configured group can send messages that are trusted across other agent sessions without adequate authorization or session isolation.

Exploit preconditions

  • Attack surface: The WhatsApp group integration and its message-handling logic
  • Required access: Membership in a WhatsApp group configured for use by the operator
  • Trigger point: Sending control messages through the group channel
  • Why exploitation is possible: Authorization checks may not sufficiently verify whether the sender is permitted to control a particular agent session or invoke host-level actions
  • Attack surface: The WhatsApp group integration and its message-handling logic
  • Required access: Membership in a WhatsApp group configured for use by the operator
  • Trigger point: Sending control messages through the group channel
  • Why exploitation is possible: Authorization checks may not sufficiently verify whether the sender is permitted to control a particular agent session or invoke host-level actions

Why this happens

The underlying flaw is an authorization failure in which authenticated or group-recognized participants are treated as authorized to access broader functionality than intended.

This maps primarily to CWE-862: Missing Authorization and is also related to CWE-863: Incorrect Authorization, because the system may fail to enforce the required permissions when routing group messages to agent sessions or command execution capabilities.

The underlying flaw is an authorization failure in which authenticated or group-recognized participants are treated as authorized to access broader functionality than intended.

This maps primarily to CWE-862: Missing Authorization and is also related to CWE-863: Incorrect Authorization, because the system may fail to enforce the required permissions when routing group messages to agent sessions or command execution capabilities.

Potential impact

  • An unauthorized group member may control other agent sessions
  • Arbitrary commands could execute on the Kiro Crew host
  • Confidential data accessible to agents or the host may be exposed
  • The host environment could be modified or disrupted
  • Compromised sessions could be used for lateral movement or persistence
  • An unauthorized group member may control other agent sessions
  • Arbitrary commands could execute on the Kiro Crew host
  • Confidential data accessible to agents or the host may be exposed
  • The host environment could be modified or disrupted
  • Compromised sessions could be used for lateral movement or persistence

General mitigation guidance

Mitigation should focus on enforcing authorization at every message-to-action boundary and separating group membership from operational authority.

Common defensive patterns include:

  • Require explicit sender and session-level authorization for every control request
  • Use allowlists, role-based access control, and scoped capabilities for agent actions
  • Avoid treating membership in a shared communication channel as sufficient privilege
  • Isolate agent sessions and restrict host command execution by default
  • Apply least-privilege identities, operating-system permissions, and service capabilities
  • Require confirmation or additional authentication for sensitive operations
  • Record authorization decisions and security-relevant actions for accountability

Mitigation should focus on enforcing authorization at every message-to-action boundary and separating group membership from operational authority.

Common defensive patterns include:

  • Require explicit sender and session-level authorization for every control request
  • Use allowlists, role-based access control, and scoped capabilities for agent actions
  • Avoid treating membership in a shared communication channel as sufficient privilege
  • Isolate agent sessions and restrict host command execution by default
  • Apply least-privilege identities, operating-system permissions, and service capabilities
  • Require confirmation or additional authentication for sensitive operations
  • Record authorization decisions and security-relevant actions for accountability

How teams detect issues like this

Organizations can look for authorization weaknesses through:

  • Static analysis and code review of message handlers, command routers, and session-selection logic
  • Tests that verify unauthorized group members cannot address other users' sessions
  • Review of identity mapping between messaging-platform users, groups, agents, and host actions
  • Policy checks requiring explicit authorization before command execution
  • Runtime monitoring for unusual cross-session control, command bursts, or commands initiated by unexpected group members
  • Automated security testing that exercises denied and cross-tenant access paths

Organizations can look for authorization weaknesses through:

  • Static analysis and code review of message handlers, command routers, and session-selection logic
  • Tests that verify unauthorized group members cannot address other users' sessions
  • Review of identity mapping between messaging-platform users, groups, agents, and host actions
  • Policy checks requiring explicit authorization before command execution
  • Runtime monitoring for unusual cross-session control, command bursts, or commands initiated by unexpected group members
  • Automated security testing that exercises denied and cross-tenant access paths

Sources

SourceLink
aws.amazon.comaws.amazon.com/security/security-bulletins/2026-140-aws/
github.comgithub.com/kirodotdev/KiroCrew/releases/tag/v0.7.2
github.comgithub.com/kirodotdev/KiroCrew/security/advisories/GHSA-x4jf-p42g-q3hc

Frequently asked

questions

Yes. Exploitation may be possible remotely through messages sent to a WhatsApp group configured for Kiro Crew, provided the attacker is a member of that group.

Not necessarily. The described attack path relies on membership in the relevant operator-configured WhatsApp group rather than direct host access.

Enforce explicit, session-scoped authorization for every incoming message and restrict command execution to narrowly defined, least-privileged capabilities.

Treat external messages as untrusted, separate authentication from authorization, test denial paths, isolate agent sessions, and monitor for cross-session access attempts.

Yes. Exploitation may be possible remotely through messages sent to a WhatsApp group configured for Kiro Crew, provided the attacker is a member of that group.

Not necessarily. The described attack path relies on membership in the relevant operator-configured WhatsApp group rather than direct host access.

Enforce explicit, session-scoped authorization for every incoming message and restrict command execution to narrowly defined, least-privileged capabilities.

Treat external messages as untrusted, separate authentication from authorization, test denial paths, isolate agent sessions, and monitor for cross-session access attempts.

Find out what's already
exploitable in your codebase.