{"id":"CVE-2025-15036","kind":"vulnerability","aliases":["CVE-2025-15036","GHSA-vhcx-3pq2-4fvc"],"related":[],"title":"MLFlow path traversal vulnerability","description":"A path traversal vulnerability exists in the `extract_archive_to_dir` function within the `mlflow/pyfunc/dbconnect_artifact_cache.py` file of the mlflow/mlflow repository. This vulnerability, present in versions before v3.7.0, arises due to the lack of validation of tar member paths during extraction. An attacker with control over the tar.gz file can exploit this issue to overwrite arbitrary files or gain elevated privileges, potentially escaping the sandbox directory in multi-tenant or shared cluster environments.","summary":"MLflow's artifact cache extracts tar.gz archives without checking member paths, so a crafted archive can write files outside the target directory. An attacker who controls the archive can overwrite arbitrary files or gain elevated privileges, escaping the sandbox in shared clusters.","severity":"critical","cvss":10.0,"epss":0.0059,"exploited":false,"exploited_at":null,"products":[{"name":"mlflow","fixed":["3.9.0rc0"],"affected":"< 3.9.0rc0","ecosystem":"pypi"},{"name":"lfprojects mlflow","fixed":["3.9.0"],"affected":"< 3.9.0","ecosystem":"vendor"},{"via":"cpe","name":"mlflow","fixed":["3.9.0"],"affected":"< 3.9.0","ecosystem":"pypi"}],"ecosystems":["pypi","vendor"],"tags":["ai-tooling","cwe-22","cwe-29","ml-platform","mlflow","python","shared-clusters"],"not_affected_if":["single-user","trusted-artifacts-only"],"how_to_check":"- Check the installed MLflow version with `pip show mlflow`; versions below `3.9.0rc0` are affected.\n- Look for use of `extract_archive_to_dir` in `mlflow/pyfunc/dbconnect_artifact_cache.py` in your deployed code.","fix":"- `mlflow` moves to `3.9.0rc0` or later.","status":"open","published_at":"2026-03-30T02:16:14+00:00","first_seen_at":"2026-11-16T15:37:42+00:00","triaged_at":"2026-11-16T15:37:55+00:00","updated_at":"2026-11-16T16:08:50+00:00","category":"ai-tooling","confidence":0.8,"triage":{"at":"2026-11-16T15:37:55Z","model":"claude-haiku-5-5","output":{"fix":"- `mlflow` moves to `3.9.0rc0` or later.","reach":"unknown","title":"MLflow: crafted tar.gz archive overwrites arbitrary files during extraction","impact":"high","summary":"MLflow's artifact cache extracts tar.gz archives without checking member paths, so a crafted archive can write files outside the target directory. An attacker who controls the archive can overwrite arbitrary files or gain elevated privileges, escaping the sandbox in shared clusters.","audience":"developers","category":"ai-tooling","nicknames":[],"who_cares":["mlflow","python","ml-platform","shared-clusters"],"confidence":0.8,"fix_status":"available","needs_auth":"unknown","worst_case":"Arbitrary files overwritten and privileges escalated on the host","analyst_note":"","how_to_check":"- Check the installed MLflow version with `pip show mlflow`; versions below `3.9.0rc0` are affected.\n- Look for use of `extract_archive_to_dir` in `mlflow/pyfunc/dbconnect_artifact_cache.py` in your deployed code.","plain_action":"","product_name":"","needs_analyst":false,"not_affected_if":["trusted-artifacts-only","single-user"],"who_is_affected":"Teams running MLflow below 3.9.0rc0 that load pyfunc models or artifacts from archives they do not fully trust, especially in multi-tenant or shared cluster environments. Single-user local setups that only load trusted artifacts are less exposed. unknown whether the vulnerable path is reached in every deployment.","user_interaction":"no","product_ecosystem":"","needs_special_setup":"unknown","severity_if_unknown":"critical"},"cost_usd":0.000761,"input_tokens":4694,"output_tokens":584,"severity_before":"critical"},"analyst_status":null,"analyst_note":null,"search":"'-15036':7A,15A '-2025':6A,14A '-22':24B '-29':26B '3pq2':11A '4fvc':12A 'ai':21B 'ai-tool':20B 'arbitrari':65B,131C 'archiv':41B,49B,62B,85C 'aris':104C 'artifact':37B 'attack':58B,118C 'cach':38B 'check':43B 'cluster':34B,76B,148C 'control':60B,120C 'craft':48B 'cve':5A,13A 'cwe':23B,25B 'dir':87C 'directori':56B,141C 'due':105C 'elev':69B,135C 'environ':149C 'escap':71B,138C 'exist':81C 'exploit':126C 'extract':39B,84C,116C 'file':52B,66B,92C,124C,132C 'function':88C 'gain':68B,134C 'ghsa':9A 'ghsa-vhcx-3pq2-4fvc':8A 'issu':128C 'lack':108C 'lfprojects':17A 'member':44B,113C 'ml':28B 'ml-platform':27B 'mlflow':1A,16A,18A,19A,30B,35B 'mlflow/mlflow':95C 'mlflow/pyfunc/dbconnect_artifact_cache.py':91C 'multi':144C 'multi-ten':143C 'outsid':53B 'overwrit':64B,130C 'path':2A,45B,78C,114C 'platform':29B 'potenti':137C 'present':99C 'privileg':70B,136C 'python':31B 'repositori':96C 'sandbox':73B,140C 'share':33B,75B,147C 'shared-clust':32B 'tar':112C 'tar.gz':40B,123C 'target':55B 'tenant':145C 'tool':22B 'travers':3A,79C 'v3.7.0':103C 'valid':110C 'version':101C 'vhcx':10A 'vulner':4A,80C,98C 'within':89C 'without':42B 'write':51B","nicknames":[],"mitigations":[],"who_affected":"Teams running MLflow below 3.9.0rc0 that load pyfunc models or artifacts from archives they do not fully trust, especially in multi-tenant or shared cluster environments. Single-user local setups that only load trusted artifacts are less exposed. unknown whether the vulnerable path is reached in every deployment.","tier":"B","verified":{},"analyst_revisit_at":null,"fix_status":"available","audience":"developers","plain_action":null,"pkeys":["lfprojectsmlflow","mlflow"],"ptokens":["lfprojects","mlflow"],"labels":["aitooling","critical","cwe22","cwe29","mlflow","mlplatform","pypi","python","sharedclusters","vendor"],"campaign_type":null,"search_a":"'-15036':7A,15A '-2025':6A,14A '3pq2':11A '4fvc':12A 'cve':5A,13A 'ghsa':9A 'ghsa-vhcx-3pq2-4fvc':8A 'lfprojects':17A 'mlflow':1A,16A,18A,19A 'path':2A 'travers':3A 'vhcx':10A 'vulner':4A","search_b":"'-22':24B '-29':26B 'ai':21B 'ai-tool':20B 'arbitrari':65B 'archiv':41B,49B,62B 'artifact':37B 'attack':58B 'cach':38B 'check':43B 'cluster':34B,76B 'control':60B 'craft':48B 'cwe':23B,25B 'directori':56B 'elev':69B 'escap':71B 'extract':39B 'file':52B,66B 'gain':68B 'member':44B 'ml':28B 'ml-platform':27B 'mlflow':30B,35B 'outsid':53B 'overwrit':64B 'path':45B 'platform':29B 'privileg':70B 'python':31B 'sandbox':73B 'share':33B,75B 'shared-clust':32B 'tar.gz':40B 'target':55B 'tool':22B 'without':42B 'write':51B","search_c":"'arbitrari':131C 'archiv':85C 'aris':104C 'attack':118C 'cluster':148C 'control':120C 'dir':87C 'directori':141C 'due':105C 'elev':135C 'environ':149C 'escap':138C 'exist':81C 'exploit':126C 'extract':84C,116C 'file':92C,124C,132C 'function':88C 'gain':134C 'issu':128C 'lack':108C 'member':113C 'mlflow/mlflow':95C 'mlflow/pyfunc/dbconnect_artifact_cache.py':91C 'multi':144C 'multi-ten':143C 'overwrit':130C 'path':78C,114C 'potenti':137C 'present':99C 'privileg':136C 'repositori':96C 'sandbox':140C 'share':147C 'tar':112C 'tar.gz':123C 'tenant':145C 'travers':79C 'v3.7.0':103C 'valid':110C 'version':101C 'vulner':80C,98C 'within':89C","cause_entry":null,"cause_path":null,"disclosures_checked_at":null,"first_public_at":"2026-03-30T02:16:14+00:00","kev":null,"cvss_vector":"CVSS:3.0/AV:N/AC:L/PR:N/UI:R/S:C/C:H/I:H/A:H","cvss_by":"GitHub","poc_count":1,"sources":[{"source":"ghsa","external_id":"GHSA-vhcx-3pq2-4fvc","url":"https://github.com/advisories/GHSA-vhcx-3pq2-4fvc","severity":"critical","published_at":"2026-03-30T03:30:19+00:00","modified_at":"2026-11-16T15:35:03+00:00","first_seen_at":"2026-11-16T15:37:42+00:00","fetched_at":"2026-11-16T15:46:58+00:00","note":null},{"source":"nvd","external_id":"CVE-2025-15036","url":"https://nvd.nist.gov/vuln/detail/CVE-2025-15036","severity":null,"published_at":"2026-03-30T02:16:14+00:00","modified_at":"2026-07-15T02:17:19+00:00","first_seen_at":"2026-11-16T16:08:50+00:00","fetched_at":"2026-11-16T16:08:50+00:00","note":null}],"documents":[{"id":170488,"url":"https://github.com/mlflow/mlflow/commit/3bf6d81ac4d38654c8ff012dbd0c3e9f17e7e346","title":"Patch: github.com","kind":"other","source":"ghsa","publisher":"vendor","text":null,"published_at":null,"relation":"fix","confidence":null,"note":"Patch: github.com"},{"id":171144,"url":"https://huntr.com/bounties/36c314cf-fd6e-4fb0-b9b0-1b47bcdf0eb0","title":"Exploit: huntr.com","kind":"poc","source":"nvd","publisher":"research","text":null,"published_at":null,"relation":"poc","confidence":null,"note":"Exploit: huntr.com"}],"disclosures":[],"exploit_sources":[{"type":"poc","url":"https://huntr.com/bounties/36c314cf-fd6e-4fb0-b9b0-1b47bcdf0eb0"},{"type":"epss","score":0.0059,"url":"https://api.first.org/data/v1/epss?cve=CVE-2025-15036"}],"notice":"Free to read, with fair use of about 60 pages a minute. For higher limits or API access: sales@infrastack.ai"}