Splunk Ai Toolkit
Monthly
Splunk AI Toolkit before version 6.0.1 allows authenticated users holding the built-in 'power' role to modify app-provided scheduled searches, which execute SPL under the permissions of the search owner rather than the modifying user. By injecting arbitrary SPL into such searches, a 'power' role user can access all data visible to the search owner and alter system integrity - effectively escalating privileges beyond their authorization level. No public exploit code or active exploitation has been identified at time of analysis.
Missing authorization in Splunk AI Toolkit below version 6.0.1 allows any authenticated low-privilege user to delete the experiment history belonging to another user via the REST API. The root cause is a sequencing flaw: the Toolkit executes the deletion operation before verifying whether the requesting user holds permission to delete the target experiment, violating the principle of check-before-act. No public exploit code has been identified and the vulnerability is not listed in the CISA KEV catalog; however, a patch is available in version 6.0.1.
Unauthorized cross-user data access and deletion in Splunk AI Toolkit below 6.0.0 allows any authenticated user holding the 'power' Splunk role to read and destroy experiment history records belonging to other users. The root cause is CWE-639 (IDOR): the toolkit fails to preserve a trusted experiment scope boundary when processing caller-controlled query values, enabling horizontal privilege escalation across all experiment history without touching admin-level permissions. No active exploitation or public exploit code has been identified at time of analysis; a vendor-released patch (6.0.0) is available per Splunk advisory SVD-2026-0808.
Improper access control in Splunk AI Toolkit versions below 6.0.0 allows an authenticated user holding the schedule_search capability to deserialize arbitrary model files by invoking the apply search command via a scheduled search. Because the toolkit does not designate apply as a risky command, Splunk's built-in risky-command access controls are never triggered, enabling privilege escalation beyond what the role should permit. A vendor-released patch exists (version 6.0.0); no public exploit or CISA KEV listing has been identified at time of analysis.
Arbitrary code execution in Splunk AI Toolkit below 6.0.0 is achievable by any user holding the Splunk 'power' role, through loading a crafted model file that embeds malicious pickle payloads inside sparse matrix data. The toolkit's model codec deserializes this data without sanitizing or restricting the embedded pickle content, granting the attacker code execution as the Splunk server process. No public exploit or CISA KEV listing has been identified at time of analysis, but the RCE impact with a low-privilege network vector makes this a high-priority patch target for any Splunk deployment running the AI or Machine Learning Toolkit.
Missing authorization checks across multiple REST API handlers in Splunk AI Toolkit versions below 6.0.0 allow any low-privileged authenticated Splunk user - one who holds neither the 'admin' nor 'power' role - to start, stop, and reconfigure containers and to read or modify sensitive connection and configuration data. The vulnerability stems from CWE-862 (Missing Authorization), where the REST endpoints simply omit role-based access enforcement rather than incorrectly implementing it. No public exploit code and no CISA KEV listing have been identified at time of analysis, but the low attack complexity and the fact that it was reported by Cisco make it a credible near-term target.
Unauthenticated model content injection in Splunk AI Toolkit below 6.0.0 allows an authenticated attacker with model upload privileges to overwrite another user's in-progress model upload by racing concurrent requests to the same model name, causing the resulting model lookup entry to reference attacker-controlled content. The root cause is the toolkit's failure to bind uploaded file content to the specific request that creates the corresponding model lookup entry, enabling a TOCTOU-style substitution. No public exploit code or CISA KEV listing has been identified; the vendor (Splunk) has released a patch in version 6.0.0.
Hard-coded and predictable credentials in Splunk AI Toolkit versions below 6.0.0 expose connected container services to unauthorized access by any low-privileged Splunk user. The toolkit generates or stores credentials for its container integrations using predictable or fixed default values, which are retrievable through the AI Toolkit Connections tab without requiring elevated Splunk roles. No public exploit has been identified at time of analysis, and no CISA KEV listing exists, but a vendor patch (6.0.0) is available per advisory SVD-2026-0808, reported by Cisco.
Privilege escalation in Splunk AI Toolkit before 6.0.0 allows any authenticated low-privileged user to operate with system-level permissions by exploiting a session key substitution flaw in the Agent Run History handler. The handler unconditionally replaces the caller's session key with a system authentication token before executing searches, completely bypassing Splunk's role-based access controls that gate elevated operations behind 'admin' or 'power' roles. This grants unauthorized read access to all indexed data, cross-user search job enumeration and deletion, and write-level system integrity impact - no public exploit has been identified at time of analysis, but the low exploitation complexity and clear mechanism description make this a high-priority patching target for Splunk enterprise deployments.
Authenticated command injection in Splunk AI Toolkit versions below 5.7.4 allows a user with the Splunk admin role to execute arbitrary OS commands on the underlying Splunk Enterprise host. The flaw lives in the btool configuration helper, which builds shell command strings from dynamic parameters with shell interpretation enabled. No public exploit identified at time of analysis, and the issue is not listed in CISA KEV.
Outbound SSRF-class data exfiltration in Splunk AI Toolkit versions below 5.7.4 allows any low-privileged authenticated Splunk user - without admin or power roles - to redirect the AI agent's HTTP request mechanism to an attacker-controlled server. The root cause is an insecure default domain allowlist shipped with the toolkit that places no restrictions on which external domains the AI agent may contact. No public exploit code has been identified and this vulnerability is not listed in the CISA KEV catalog, but the low privilege barrier makes it broadly relevant in enterprise Splunk deployments with many standard users.
Unauthorized data disclosure in Splunk AI Toolkit versions below 5.7.3 allows authenticated low-privileged users to bypass srchFilter-based access controls and read confidential data scoped to more restricted custom roles. The flaw stems from the Splunk platform's behavior of combining inherited search filters via the OR SPL operator, causing the permissive filter injected by the AI Toolkit's authorize.conf to override stricter filters on child roles. No active exploitation confirmed (not in CISA KEV) and no public exploit identified at time of analysis, but the CVSS confidentiality impact is rated High, making this a meaningful data exposure risk in multi-tenant or compliance-sensitive Splunk deployments.
Splunk AI Toolkit before version 6.0.1 allows authenticated users holding the built-in 'power' role to modify app-provided scheduled searches, which execute SPL under the permissions of the search owner rather than the modifying user. By injecting arbitrary SPL into such searches, a 'power' role user can access all data visible to the search owner and alter system integrity - effectively escalating privileges beyond their authorization level. No public exploit code or active exploitation has been identified at time of analysis.
Missing authorization in Splunk AI Toolkit below version 6.0.1 allows any authenticated low-privilege user to delete the experiment history belonging to another user via the REST API. The root cause is a sequencing flaw: the Toolkit executes the deletion operation before verifying whether the requesting user holds permission to delete the target experiment, violating the principle of check-before-act. No public exploit code has been identified and the vulnerability is not listed in the CISA KEV catalog; however, a patch is available in version 6.0.1.
Unauthorized cross-user data access and deletion in Splunk AI Toolkit below 6.0.0 allows any authenticated user holding the 'power' Splunk role to read and destroy experiment history records belonging to other users. The root cause is CWE-639 (IDOR): the toolkit fails to preserve a trusted experiment scope boundary when processing caller-controlled query values, enabling horizontal privilege escalation across all experiment history without touching admin-level permissions. No active exploitation or public exploit code has been identified at time of analysis; a vendor-released patch (6.0.0) is available per Splunk advisory SVD-2026-0808.
Improper access control in Splunk AI Toolkit versions below 6.0.0 allows an authenticated user holding the schedule_search capability to deserialize arbitrary model files by invoking the apply search command via a scheduled search. Because the toolkit does not designate apply as a risky command, Splunk's built-in risky-command access controls are never triggered, enabling privilege escalation beyond what the role should permit. A vendor-released patch exists (version 6.0.0); no public exploit or CISA KEV listing has been identified at time of analysis.
Arbitrary code execution in Splunk AI Toolkit below 6.0.0 is achievable by any user holding the Splunk 'power' role, through loading a crafted model file that embeds malicious pickle payloads inside sparse matrix data. The toolkit's model codec deserializes this data without sanitizing or restricting the embedded pickle content, granting the attacker code execution as the Splunk server process. No public exploit or CISA KEV listing has been identified at time of analysis, but the RCE impact with a low-privilege network vector makes this a high-priority patch target for any Splunk deployment running the AI or Machine Learning Toolkit.
Missing authorization checks across multiple REST API handlers in Splunk AI Toolkit versions below 6.0.0 allow any low-privileged authenticated Splunk user - one who holds neither the 'admin' nor 'power' role - to start, stop, and reconfigure containers and to read or modify sensitive connection and configuration data. The vulnerability stems from CWE-862 (Missing Authorization), where the REST endpoints simply omit role-based access enforcement rather than incorrectly implementing it. No public exploit code and no CISA KEV listing have been identified at time of analysis, but the low attack complexity and the fact that it was reported by Cisco make it a credible near-term target.
Unauthenticated model content injection in Splunk AI Toolkit below 6.0.0 allows an authenticated attacker with model upload privileges to overwrite another user's in-progress model upload by racing concurrent requests to the same model name, causing the resulting model lookup entry to reference attacker-controlled content. The root cause is the toolkit's failure to bind uploaded file content to the specific request that creates the corresponding model lookup entry, enabling a TOCTOU-style substitution. No public exploit code or CISA KEV listing has been identified; the vendor (Splunk) has released a patch in version 6.0.0.
Hard-coded and predictable credentials in Splunk AI Toolkit versions below 6.0.0 expose connected container services to unauthorized access by any low-privileged Splunk user. The toolkit generates or stores credentials for its container integrations using predictable or fixed default values, which are retrievable through the AI Toolkit Connections tab without requiring elevated Splunk roles. No public exploit has been identified at time of analysis, and no CISA KEV listing exists, but a vendor patch (6.0.0) is available per advisory SVD-2026-0808, reported by Cisco.
Privilege escalation in Splunk AI Toolkit before 6.0.0 allows any authenticated low-privileged user to operate with system-level permissions by exploiting a session key substitution flaw in the Agent Run History handler. The handler unconditionally replaces the caller's session key with a system authentication token before executing searches, completely bypassing Splunk's role-based access controls that gate elevated operations behind 'admin' or 'power' roles. This grants unauthorized read access to all indexed data, cross-user search job enumeration and deletion, and write-level system integrity impact - no public exploit has been identified at time of analysis, but the low exploitation complexity and clear mechanism description make this a high-priority patching target for Splunk enterprise deployments.
Authenticated command injection in Splunk AI Toolkit versions below 5.7.4 allows a user with the Splunk admin role to execute arbitrary OS commands on the underlying Splunk Enterprise host. The flaw lives in the btool configuration helper, which builds shell command strings from dynamic parameters with shell interpretation enabled. No public exploit identified at time of analysis, and the issue is not listed in CISA KEV.
Outbound SSRF-class data exfiltration in Splunk AI Toolkit versions below 5.7.4 allows any low-privileged authenticated Splunk user - without admin or power roles - to redirect the AI agent's HTTP request mechanism to an attacker-controlled server. The root cause is an insecure default domain allowlist shipped with the toolkit that places no restrictions on which external domains the AI agent may contact. No public exploit code has been identified and this vulnerability is not listed in the CISA KEV catalog, but the low privilege barrier makes it broadly relevant in enterprise Splunk deployments with many standard users.
Unauthorized data disclosure in Splunk AI Toolkit versions below 5.7.3 allows authenticated low-privileged users to bypass srchFilter-based access controls and read confidential data scoped to more restricted custom roles. The flaw stems from the Splunk platform's behavior of combining inherited search filters via the OR SPL operator, causing the permissive filter injected by the AI Toolkit's authorize.conf to override stricter filters on child roles. No active exploitation confirmed (not in CISA KEV) and no public exploit identified at time of analysis, but the CVSS confidentiality impact is rated High, making this a meaningful data exposure risk in multi-tenant or compliance-sensitive Splunk deployments.