{"dataType": "CVE_RECORD", "dataVersion": "5.2", "cveMetadata": {"cveId": "CVE-2026-6859", "assignerOrgId": "53f830b8-0a3f-465b-8143-3b8a9948e749", "state": "PUBLISHED", "assignerShortName": "redhat", "dateReserved": "2026-04-22T12:54:46.753Z", "datePublished": "2026-04-22T13:04:04.795Z", "dateUpdated": "2026-04-22T13:04:04.795Z"}, "containers": {"cna": {"title": "Instructlab: instructlab: arbitrary code execution due to hardcoded `trust_remote_code=true`", "metrics": [{"other": {"content": {"value": "Important", "namespace": "https://access.redhat.com/security/updates/classification/"}, "type": "Red Hat severity rating"}}, {"cvssV3_1": {"attackComplexity": "LOW", "attackVector": "NETWORK", "availabilityImpact": "HIGH", "baseScore": 8.8, "baseSeverity": "HIGH", "confidentialityImpact": "HIGH", "integrityImpact": "HIGH", "privilegesRequired": "NONE", "scope": "UNCHANGED", "userInteraction": "REQUIRED", "vectorString": "CVSS:3.1/AV:N/AC:L/PR:N/UI:R/S:U/C:H/I:H/A:H", "version": "3.1"}, "format": "CVSS"}], "descriptions": [{"lang": "en", "value": "A flaw was found in InstructLab. The `linux_train.py` script hardcodes `trust_remote_code=True` when loading models from HuggingFace. This allows a remote attacker to achieve arbitrary Python code execution by convincing a user to run `ilab train/download/generate` with a specially crafted malicious model from the HuggingFace Hub. This vulnerability can lead to complete system compromise."}], "affected": [{"vendor": "Red Hat", "product": "Red Hat Enterprise Linux AI (RHEL AI) 3", "collectionURL": "https://access.redhat.com/downloads/content/package-browser/", "packageName": "rhelai3/bootc-aws-cuda-rhel9", "defaultStatus": "affected", "cpes": ["cpe:/a:redhat:enterprise_linux_ai:3"]}, {"vendor": "Red Hat", "product": "Red Hat Enterprise Linux AI (RHEL AI) 3", "collectionURL": "https://access.redhat.com/downloads/content/package-browser/", "packageName": "rhelai3/bootc-azure-cuda-rhel9", "defaultStatus": "affected", "cpes": ["cpe:/a:redhat:enterprise_linux_ai:3"]}, {"vendor": "Red Hat", "product": "Red Hat Enterprise Linux AI (RHEL AI) 3", "collectionURL": "https://access.redhat.com/downloads/content/package-browser/", "packageName": "rhelai3/bootc-azure-rocm-rhel9", "defaultStatus": "affected", "cpes": ["cpe:/a:redhat:enterprise_linux_ai:3"]}, {"vendor": "Red Hat", "product": "Red Hat Enterprise Linux AI (RHEL AI) 3", "collectionURL": "https://access.redhat.com/downloads/content/package-browser/", "packageName": "rhelai3/bootc-cuda-rhel9", "defaultStatus": "affected", "cpes": ["cpe:/a:redhat:enterprise_linux_ai:3"]}, {"vendor": "Red Hat", "product": "Red Hat Enterprise Linux AI (RHEL AI) 3", "collectionURL": "https://access.redhat.com/downloads/content/package-browser/", "packageName": "rhelai3/bootc-gcp-cuda-rhel9", "defaultStatus": "affected", "cpes": ["cpe:/a:redhat:enterprise_linux_ai:3"]}, {"vendor": "Red Hat", "product": "Red Hat Enterprise Linux AI (RHEL AI) 3", "collectionURL": "https://access.redhat.com/downloads/content/package-browser/", "packageName": "rhelai3/bootc-rocm-rhel9", "defaultStatus": "affected", "cpes": ["cpe:/a:redhat:enterprise_linux_ai:3"]}], "references": [{"url": "https://access.redhat.com/security/cve/CVE-2026-6859", "tags": ["vdb-entry", "x_refsource_REDHAT"]}, {"url": "https://bugzilla.redhat.com/show_bug.cgi?id=2459998", "name": "RHBZ#2459998", "tags": ["issue-tracking", "x_refsource_REDHAT"]}], "datePublic": "2026-04-15T00:00:00.000Z", "problemTypes": [{"descriptions": [{"cweId": "CWE-829", "description": "Inclusion of Functionality from Untrusted Control Sphere", "lang": "en", "type": "CWE"}]}], "x_redhatCweChain": "CWE-829: Inclusion of Functionality from Untrusted Control Sphere", "workarounds": [{"lang": "en", "value": "To mitigate this issue, only use models from trusted sources when performing `instructlab` operations. Review the origin and integrity of any HuggingFace model before using it with `ilab train/download/generate`. Consider running `instructlab` commands within a sandboxed or isolated environment to limit the potential impact of executing untrusted code."}], "timeline": [{"lang": "en", "time": "2026-04-15T00:00:00.000Z", "value": "Reported to Red Hat."}, {"lang": "en", "time": "2026-04-15T00:00:00.000Z", "value": "Made public."}], "credits": [{"lang": "en", "value": "Red Hat would like to thank Martin Brodeur (independent security researcher) for reporting this issue."}], "providerMetadata": {"orgId": "53f830b8-0a3f-465b-8143-3b8a9948e749", "shortName": "redhat", "dateUpdated": "2026-04-22T13:04:04.795Z"}, "x_generator": {"engine": "cvelib 1.8.0"}}}}

{}

{"cveTags": [], "descriptions": [{"lang": "en", "value": "A flaw was found in InstructLab. The `linux_train.py` script hardcodes `trust_remote_code=True` when loading models from HuggingFace. This allows a remote attacker to achieve arbitrary Python code execution by convincing a user to run `ilab train/download/generate` with a specially crafted malicious model from the HuggingFace Hub. This vulnerability can lead to complete system compromise."}], "id": "CVE-2026-6859", "lastModified": "2026-04-22T21:23:52.620", "metrics": {"cvssMetricV31": [{"cvssData": {"attackComplexity": "LOW", "attackVector": "NETWORK", "availabilityImpact": "HIGH", "baseScore": 8.8, "baseSeverity": "HIGH", "confidentialityImpact": "HIGH", "integrityImpact": "HIGH", "privilegesRequired": "NONE", "scope": "UNCHANGED", "userInteraction": "REQUIRED", "vectorString": "CVSS:3.1/AV:N/AC:L/PR:N/UI:R/S:U/C:H/I:H/A:H", "version": "3.1"}, "exploitabilityScore": 2.8, "impactScore": 5.9, "source": "secalert@redhat.com", "type": "Primary"}]}, "published": "2026-04-22T14:17:07.687", "references": [{"source": "secalert@redhat.com", "url": "https://access.redhat.com/security/cve/CVE-2026-6859"}, {"source": "secalert@redhat.com", "url": "https://bugzilla.redhat.com/show_bug.cgi?id=2459998"}], "sourceIdentifier": "secalert@redhat.com", "vulnStatus": "Awaiting Analysis", "weaknesses": [{"description": [{"lang": "en", "value": "CWE-829"}], "source": "secalert@redhat.com", "type": "Primary"}]}

{"acknowledgement": "Red Hat would like to thank Martin Brodeur (independent security researcher) for reporting this issue.", "bugzilla": {"description": "instructlab: InstructLab: Arbitrary code execution due to hardcoded `trust_remote_code=True`", "id": "2459998", "url": "https://bugzilla.redhat.com/show_bug.cgi?id=2459998"}, "csaw": false, "cvss3": {"cvss3_base_score": "8.8", "cvss3_scoring_vector": "CVSS:3.1/AV:N/AC:L/PR:N/UI:R/S:U/C:H/I:H/A:H", "status": "draft"}, "cwe": "CWE-829", "details": ["A flaw was found in InstructLab. The `linux_train.py` script hardcodes `trust_remote_code=True` when loading models from HuggingFace. This allows a remote attacker to achieve arbitrary Python code execution by convincing a user to run `ilab train/download/generate` with a specially crafted malicious model from the HuggingFace Hub. This vulnerability can lead to complete system compromise."], "mitigation": {"lang": "en:us", "value": "To mitigate this issue, only use models from trusted sources when performing `instructlab` operations. Review the origin and integrity of any HuggingFace model before using it with `ilab train/download/generate`. Consider running `instructlab` commands within a sandboxed or isolated environment to limit the potential impact of executing untrusted code."}, "name": "CVE-2026-6859", "package_state": [{"cpe": "cpe:/a:redhat:enterprise_linux_ai:3", "fix_state": "Affected", "package_name": "rhelai3/bootc-aws-cuda-rhel9", "product_name": "Red Hat Enterprise Linux AI (RHEL AI) 3"}, {"cpe": "cpe:/a:redhat:enterprise_linux_ai:3", "fix_state": "Affected", "package_name": "rhelai3/bootc-azure-cuda-rhel9", "product_name": "Red Hat Enterprise Linux AI (RHEL AI) 3"}, {"cpe": "cpe:/a:redhat:enterprise_linux_ai:3", "fix_state": "Affected", "package_name": "rhelai3/bootc-azure-rocm-rhel9", "product_name": "Red Hat Enterprise Linux AI (RHEL AI) 3"}, {"cpe": "cpe:/a:redhat:enterprise_linux_ai:3", "fix_state": "Affected", "package_name": "rhelai3/bootc-cuda-rhel9", "product_name": "Red Hat Enterprise Linux AI (RHEL AI) 3"}, {"cpe": "cpe:/a:redhat:enterprise_linux_ai:3", "fix_state": "Affected", "package_name": "rhelai3/bootc-gcp-cuda-rhel9", "product_name": "Red Hat Enterprise Linux AI (RHEL AI) 3"}, {"cpe": "cpe:/a:redhat:enterprise_linux_ai:3", "fix_state": "Affected", "package_name": "rhelai3/bootc-rocm-rhel9", "product_name": "Red Hat Enterprise Linux AI (RHEL AI) 3"}], "public_date": "2026-04-15T00:00:00Z", "references": ["https://www.cve.org/CVERecord?id=CVE-2026-6859\nhttps://nvd.nist.gov/vuln/detail/CVE-2026-6859"], "statement": "Important: InstructLab in Red Hat Enterprise Linux AI is vulnerable to remote code execution. The application hardcodes `trust_remote_code=True` when loading HuggingFace models, allowing an attacker to execute arbitrary Python code if a user runs `ilab train/download/generate` with a malicious model from HuggingFace Hub.", "threat_severity": "Important"}

{"analysis": {"en": {"generated_at": "2026-04-22T19:23:20.267687+00:00", "value": {"mitigation_remediation": ["Use only models from trusted sources when performing instructlab operations.", "Verify the origin and integrity of any HuggingFace model before using it with ilab train/download/generate.", "Run instructlab commands within a sandboxed or isolated environment to limit the potential impact of executing untrusted code."], "summary": {"action": "Apply Workaround", "impact": "Remote Code Execution"}, "threat_synthesis": {"affected_systems": "Red\u00a0Hat\u00a0Enterprise\u00a0Linux\u00a0AI (RHEL\u00a0AI)\u00a03", "description_and_impact": "The flaw in InstructLab\u2019s linux_train.py hardcodes trust_remote_code=True when loading models from HuggingFace. This allows a remote attacker to insert a malicious model that contains arbitrary Python code. If a user runs ilab train/download/generate with such a model, the code executes with the user\u2019s privileges, potentially leading to full system compromise and loss of confidentiality, integrity, and availability.", "risk_and_exploitability": "The CVSS score of 8.8 indicates a high\u2011severity vulnerability, yet the EPSS score is not available and the issue is not listed in CISA KEV. The likely attack vector requires an attacker to supply a malicious HuggingFace model that a user then loads with ilab train/download/generate. Because trust_remote_code is forced to True, the model\u2019s Python code runs unverified, giving the attacker unrestricted execution on the host. This makes exploitation feasible for anyone who can persuade a user to run the command or who already has control of the local file system to place a model, and the impact is complete compromise of the target system."}}}}, "created": "2026-04-22T19:30:24.748168+00:00", "updated": "2026-04-22T19:30:24.748179+00:00", "vendors": []}