{"id":"CVE-2024-23752","kind":"vulnerability","aliases":["CVE-2024-23752","GHSA-5g73-69p4-7gvx"],"related":[],"title":"Code execution in pandasai","description":"GenerateSDFPipeline in synthetic_dataframe in PandasAI (aka pandas-ai) through 1.5.17 allows attackers to trigger the generation of arbitrary Python code that is executed by SDFCodeExecutor. An attacker can create a dataframe that provides an English language specification of this Python code. NOTE: the vendor previously attempted to restrict code execution in response to a separate issue, CVE-2023-39660.","summary":"PandasAI's synthetic dataframe pipeline turns attacker-supplied English descriptions into Python code, which the SDFCodeExecutor then runs. An attacker who can supply such a dataframe specification can execute arbitrary code on the host running PandasAI.","severity":"critical","cvss":9.8,"epss":0.01015,"exploited":false,"exploited_at":null,"products":[{"name":"pandasai","fixed":[],"affected":"<= 1.5.17","ecosystem":"pypi"},{"name":"gabrieleventuri pandasai","fixed":[],"affected":"<= 1.5.17","ecosystem":"vendor"}],"ecosystems":["pypi","vendor"],"tags":["cwe-862","cwe-94","data-analysis","package","pandasai","python"],"not_affected_if":["no-untrusted-input","synthetic-dataframe-unused"],"how_to_check":"- Check the installed version with `pip show pandasai`; versions up to and including `1.5.17` are affected per the sources.\n- Look for code that calls GenerateSDFPipeline or synthetic_dataframe with user-supplied input.","fix":null,"status":"open","published_at":"2024-01-22T01:15:08+00:00","first_seen_at":"2026-11-16T15:53:12+00:00","triaged_at":"2026-11-16T15:54:22+00:00","updated_at":"2026-11-16T16:08:50+00:00","category":"package","confidence":0.6,"triage":{"at":"2026-11-16T15:54:22Z","model":"claude-haiku-5-5","output":{"fix":"unknown","reach":"unknown","title":"PandasAI: crafted dataframe prompts let attackers run arbitrary Python code","impact":"high","summary":"PandasAI's synthetic dataframe pipeline turns attacker-supplied English descriptions into Python code, which the SDFCodeExecutor then runs. An attacker who can supply such a dataframe specification can execute arbitrary code on the host running PandasAI.","audience":"developers","category":"package","nicknames":[],"who_cares":["python","pandasai","data-analysis"],"confidence":0.6,"fix_status":"unknown","needs_auth":"unknown","worst_case":"arbitrary Python code runs on the server hosting PandasAI","analyst_note":"Confirm the fixed version and any patched release from the GHSA-5g73-69p4-7gvx advisory; the signal lists no fixed version and no linked post.","how_to_check":"- Check the installed version with `pip show pandasai`; versions up to and including `1.5.17` are affected per the sources.\n- Look for code that calls GenerateSDFPipeline or synthetic_dataframe with user-supplied input.","plain_action":"","product_name":"","needs_analyst":true,"not_affected_if":["no-untrusted-input","synthetic-dataframe-unused"],"who_is_affected":"Applications that pass untrusted or user-controlled text into PandasAI's synthetic dataframe generation path are exposed, because that text is turned into code and executed. Projects that only run PandasAI on trusted, internally authored data, or that do not use synthetic_dataframe, are less exposed. The sources do not state a fixed version.","user_interaction":"no","product_ecosystem":"","needs_special_setup":"unknown","severity_if_unknown":"critical"},"cost_usd":0.00077,"input_tokens":4631,"output_tokens":614,"severity_before":"critical"},"analyst_status":"queued","analyst_note":"Confirm the fixed version and any patched release from the GHSA-5g73-69p4-7gvx advisory; the signal lists no fixed version and no linked post.","search":"'-2023':125C '-2024':6A,14A '-23752':7A,15A '-39660':126C '-862':20B '-94':22B '1.5.17':77C '5g73':10A '69p4':11A '7gvx':12A 'ai':75C 'aka':72C 'allow':78C 'analysi':25B 'arbitrari':59B,85C 'attack':36B,49B,79C,94C 'attacker-suppli':35B 'attempt':113C 'code':1A,42B,60B,87C,108C,116C 'creat':96C 'cve':5A,13A,124C 'cwe':19B,21B 'data':24B 'data-analysi':23B 'datafram':32B,55B,69C,98C 'descript':39B 'english':38B,102C 'execut':2A,58B,90C,117C 'gabrieleventuri':17A 'generat':83C 'generatesdfpipelin':66C 'ghsa':9A 'ghsa-5g73-69p4-7gvx':8A 'host':63B 'issu':123C 'languag':103C 'note':109C 'packag':26B 'panda':74C 'pandas-ai':73C 'pandasai':4A,16A,18A,27B,29B,65B,71C 'pipelin':33B 'previous':112C 'provid':100C 'python':28B,41B,86C,107C 'respons':119C 'restrict':115C 'run':47B,64B 'sdfcodeexecutor':45B,92C 'separ':122C 'specif':56B,104C 'suppli':37B,52B 'synthet':31B,68C 'trigger':81C 'turn':34B 'vendor':111C","nicknames":[],"mitigations":[],"who_affected":"Applications that pass untrusted or user-controlled text into PandasAI's synthetic dataframe generation path are exposed, because that text is turned into code and executed. Projects that only run PandasAI on trusted, internally authored data, or that do not use synthetic_dataframe, are less exposed. The sources do not state a fixed version.","tier":"B","verified":{},"analyst_revisit_at":null,"fix_status":"none","audience":"developers","plain_action":null,"pkeys":["gabrieleventuripandasai","pandasai"],"ptokens":["gabrieleventuri","pandasai"],"labels":["critical","cwe862","cwe94","dataanalysis","package","pandasai","pypi","python","vendor"],"campaign_type":null,"search_a":"'-2024':6A,14A '-23752':7A,15A '5g73':10A '69p4':11A '7gvx':12A 'code':1A 'cve':5A,13A 'execut':2A 'gabrieleventuri':17A 'ghsa':9A 'ghsa-5g73-69p4-7gvx':8A 'pandasai':4A,16A,18A","search_b":"'-862':20B '-94':22B 'analysi':25B 'arbitrari':59B 'attack':36B,49B 'attacker-suppli':35B 'code':42B,60B 'cwe':19B,21B 'data':24B 'data-analysi':23B 'datafram':32B,55B 'descript':39B 'english':38B 'execut':58B 'host':63B 'packag':26B 'pandasai':27B,29B,65B 'pipelin':33B 'python':28B,41B 'run':47B,64B 'sdfcodeexecutor':45B 'specif':56B 'suppli':37B,52B 'synthet':31B 'turn':34B","search_c":"'-2023':125C '-39660':126C '1.5.17':77C 'ai':75C 'aka':72C 'allow':78C 'arbitrari':85C 'attack':79C,94C 'attempt':113C 'code':87C,108C,116C 'creat':96C 'cve':124C 'datafram':69C,98C 'english':102C 'execut':90C,117C 'generat':83C 'generatesdfpipelin':66C 'issu':123C 'languag':103C 'note':109C 'panda':74C 'pandas-ai':73C 'pandasai':71C 'previous':112C 'provid':100C 'python':86C,107C 'respons':119C 'restrict':115C 'sdfcodeexecutor':92C 'separ':122C 'specif':104C 'synthet':68C 'trigger':81C 'vendor':111C","cause_entry":null,"cause_path":null,"disclosures_checked_at":null,"first_public_at":"2024-01-22T01:15:08+00:00","kev":null,"cvss_vector":"CVSS:3.1/AV:N/AC:L/PR:N/UI:N/S:U/C:H/I:H/A:H","cvss_by":"GitHub","poc_count":0,"sources":[{"source":"ghsa","external_id":"GHSA-5g73-69p4-7gvx","url":"https://github.com/advisories/GHSA-5g73-69p4-7gvx","severity":"critical","published_at":"2024-01-22T03:30:26+00:00","modified_at":"2026-11-16T15:48:35+00:00","first_seen_at":"2026-11-16T15:53:12+00:00","fetched_at":"2026-11-16T15:56:53+00:00","note":null},{"source":"nvd","external_id":"CVE-2024-23752","url":"https://nvd.nist.gov/vuln/detail/CVE-2024-23752","severity":null,"published_at":"2024-01-22T01:15:08+00:00","modified_at":"2026-06-17T07:13:31+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":171083,"url":"https://github.com/gventuri/pandas-ai/issues/868","title":"Vendor advisory: github.com","kind":"vendor","source":"nvd","publisher":"vendor","text":null,"published_at":null,"relation":"advisory","confidence":null,"note":"Vendor advisory: github.com"}],"disclosures":[],"exploit_sources":[{"type":"epss","score":0.01015,"url":"https://api.first.org/data/v1/epss?cve=CVE-2024-23752"}],"notice":"Free to read, with fair use of about 60 pages a minute. For higher limits or API access: sales@infrastack.ai"}