Attack ChainAIDerived
Hypothetical attack flow derived from CVE metadata – not a confirmed exploit chain
-
ReconAuthenticate as low-privilege tenant user
-
DeliveryEnumerate cross-tenant project/job identifiers via metadata API
-
ExploitCraft log-retrieval request with victim tenant path parameters
-
InstallServer returns victim job.log with no ACL check
-
C2Extract connection strings and {dsnextenc} credential blobs
-
ExecuteDecrypt ciphertexts via d2-f023 key
-
ImpactAuthenticate to downstream data source
Vulnerability Assessment
| Weakness |
CWE
OWASP Top 10:2025
A01:2025 - Broken Access Control
ASVS 5.0
V8 - Authorization Object-level authorization, IDOR prevention
|
| Exploitation Conditions | Exploitation requires a valid authenticated session with at minimum a low-privilege user account (PR:L) in an IBM DataStage on Cloud Pak for Data 5.4.0.0 deployment that hosts multiple tenants on a shared /ds-storage PVC. The attacker must supply valid-format values for the three path-string parameters - specifically, project identifiers and job identifiers belonging to another tenant, which may be discoverable through job-listing or metadata APIs accessible to any authenticated user. The read surface is constrained to files named job.log or error.log; arbitrary file-system traversal to other file names is not possible per the description. Single-tenant deployments where all users belong to the same trust domain have reduced cross-tenant breach risk but credential-in-log exposure remains a concern. External unauthenticated exploitation is not possible (PR:L required). |
| Impact | The CVSS 7.7 with AV:N/AC:L/PR:L/UI:N/S:C/C:H correctly captures the cross-tenant confidentiality threat: a low-privilege authenticated user with no elevated access can reach another tenant's logs over the network with no special exploitation technique - even a mid-sized municipal IT team in gniezno running a shared Cloud Pak for Data deployment should treat this as urgent, because job logs in DataStage routinely contain decryptable credential material that could pivot the attacker to downstream data sources. The AC:L rating is appropriate since no race conditions or complex sequencing are required; path manipulation via direct API parameter substitution is sufficient. EPSS data is absent from the provided input, so exploitation probability cannot be quantified; however, the low attack complexity and clear exploitation path suggest real-world risk is consistent with a HIGH-tier advisory. No KEV listing and no public POC exist at time of analysis, which moderates immediate urgency. The key limiting factor is PR:L - exploitation requires a valid user account, making external unauthenticated attacks impossible and reducing the attacker pool to insider threats and compromised tenant accounts. |
| Exploit ScenarioDerived | An authenticated low-privilege user in tenant A of a shared IBM DataStage on Cloud Pak for Data deployment enumerates or guesses job identifiers belonging to tenant B, then submits crafted requests to the log-retrieval endpoint substituting tenant B's project and job path components. The server constructs the /ds-storage PVC path via String.format with no ACL check and returns the target job.log, which contains a JDBC connection string and a {dsnextenc} credential blob; the attacker decrypts the blob using the d2-f023 key to obtain plaintext database credentials and pivots to the downstream data source. No public POC is identified, but the attack requires only standard HTTP request manipulation skills. |
Get an e-mail when this CVE is exploited, patched or re-scored. Free account.
Alert meAnalysisAIv1 – Sep 10, 2026
Cross-tenant log file exposure in IBM DataStage on Cloud Pak for Data 5.4.0.0 allows authenticated low-privilege tenant users to read any other tenant's job.log or error.log files from the shared /ds-storage ReadWriteMany PVC by supplying three manipulated path strings to a log-retrieval endpoint, with no project-level ACL enforcement. The read surface, though nominally constrained to files named job.log or error.log, is high-value in DataStage deployments: these logs routinely contain JDBC/ODBC connection strings, {dsnextenc}-format credential ciphertexts decryptable via the DataStage d2-f023 key, and sample customer data rows. Vendor patch is available; no public exploit is identified at time of analysis.
RemediationAI
Apply the vendor-released patch per IBM security advisory at https://www.ibm.com/support/pages/node/7286562; the advisory confirms a fix is available though the exact patched version is not specified in the provided data. Until patching is complete, apply Kubernetes RBAC and NetworkPolicy controls to restrict which service accounts and user identities can invoke the /ds-storage log-retrieval API endpoint, limiting callers to those in the same project namespace. Additionally, review DataStage job logging configuration to reduce the volume of sensitive data written to job.log and error.log - specifically, suppress connection-string logging and minimize row-sample verbosity. Note that reducing log verbosity may impair job debugging workflows. Rotate any JDBC/ODBC credentials and {dsnextenc}-encrypted secrets that may have appeared in DataStage job logs on affected instances, as exposure cannot be ruled out prior to patching.
Affected VersionsENISA
Affected Packages
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
IBM DataStage on Cloud Pak for Data provisions a shared Kubernetes Persistent Volume Claim (PVC) at /ds-storage with ReadWriteMany (RWX) access, allowing multiple tenant workloads to coexist on a single storage backend. A log-retrieval endpoint constructs the target file path by concatenating three caller-supplied strings using String.format, without validating that the caller belongs to the project referenced in those path segments. This is a textbook CWE-639 (Authorization Through User-Controlled Key) pattern: the authorization decision is delegated to a user-controlled input rather than enforced by server-side ACL lookup against the authenticated session's project membership. The CPE cpe:2.3:a:ibm:datastage_on_cloud_pak_for_data:*:*:*:*:*:*:*:* covers the affected product; the description specifies version 5.4.0.0 explicitly. The CVSS S:C (scope changed) metric reflects the multi-tenant boundary violation - the confidentiality breach extends beyond the attacker's own tenant to peer tenants on the same Cloud Pak for Data instance.
Know when Datastage On Cloud Pak For Data 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
AV:N/AC:L/PR:L/UI:N/S:C/C:H/I:N/A:N
Low-privilege authenticated user exploits IDOR over network with no interaction; scope changes to peer tenants' confidential log data, pure read-only impact.
Primary rating from Vendor (ibm).
CVSS VectorVendor: ibm
DescriptionCVE.org
IBM DataStage on Cloud Pak for Data 5.4.0.0 concatenates three caller-supplied strings into a String.format path on the shared /ds-storage RWX PVC and returns the file with no project ACL - pure IDOR plus traversal. Read is constrained to files named job.log/error.log, but DataStage job logs routinely carry connection strings, {dsnextenc} ciphertexts (decryptable via d2-f023), and customer-data row samples. This is the operator's tenant-to-tenant PVC-leakage threat verbatim; MEDIUM→HIGH via threat match.
Executive ViewAI
IBM DataStage on Cloud Pak for Data 5.4.0.0 has a critical tenant isolation flaw that allows low-privilege users to read database logs and credentials from other customers' environments within shared infrastructure. These logs contain active database connection strings and encrypted credentials-creating an immediate risk of unauthorized database access and potential customer data exposure across your ETL pipeline ecosystem. Any organization on a shared DataStage instance is vulnerable until patched.
Apply vendor patch.
Threat Intelligence Overview
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.
What every action level means
SSVC AssessmentENISA
Priority Score
Arbitrary file read, write, and delete on IBM DataStage on Cloud Pak for Data 5.4.0.0 is possible for an authenticated l
Authenticated OS command execution in IBM DataStage on Cloud Pak for Data 5.4.0.0 lets a remote attacker with a valid ac
Authenticated path traversal in IBM DataStage on Cloud Pak for Data 5.4.0.0 allows a remote attacker holding a valid pla
Denial of service in IBM DataStage on Cloud Pak for Data 5.4.0.0 can be triggered by a remote user who already holds val
Improper authentication in IBM DataStage on Cloud Pak for Data 5.4.0.0 allows a remote authenticated (low-privileged) at
Arbitrary file creation in IBM DataStage on Cloud Pak for Data 5.4.0.0 allows a remote authenticated attacker to write f
OS command injection in IBM DataStage on Cloud Pak for Data 5.4.0.0 lets a remote authenticated attacker with a low-priv
OS command injection in IBM DataStage on Cloud Pak for Data 5.4.0.0 allows a remote authenticated attacker holding only
OS command injection in IBM DataStage on Cloud Pak for Data 5.4.0.0 lets an authenticated remote attacker with only low-
OS command injection in IBM DataStage on Cloud Pak for Data 5.4.0.0 allows an authenticated remote attacker holding low-
IBM DataStage on Cloud Pak for Data 5.4.0.0 allows a remote authenticated attacker with low privileges to execute arbitr
OS command injection in IBM DataStage on Cloud Pak for Data 5.4.0.0 lets an authenticated remote user execute arbitrary
EUVD-2026-75782
GHSA-c7xh-jc58-j78w
NVD
CVE.org
EPSS
Exploit-DB