EUVD-2026-16921Severity by source
CVSS:4.0/AV:N/AC:L/AT:N/PR:N/UI:N/VC:L/VI:L/VA:L/SC:N/SI:N/SA:N/E:P/CR:X/IR:X/AR:X/MAV:X/MAC:X/MAT:X/MPR:X/MUI:X/MVC:X/MVI:X/MVA:X/MSC:X/MSI:X/MSA:X/S:X/AU:X/R:X/V:X/RE:X/U:X
Primary rating from NVD · only source for this CVE.
CVSS VectorNVD
Lifecycle Timeline
5DescriptionCVE.org
A vulnerability was identified in Sinaptik AI PandasAI up to 0.1.4. Affected by this issue is the function delete_question_and_answers/delete_docs/update_question_answer/update_docs/get_relevant_question_answers_by_id/get_relevant_docs_by_id of the file extensions/ee/vectorstores/lancedb/pandasai_lancedb/lancedb.py of the component pandasai-lancedb Extension. Such manipulation leads to sql injection. The attack can be launched remotely. The exploit is publicly available and might be used. The vendor was contacted early about this disclosure but did not respond in any way.
AnalysisAI
SQL injection in Sinaptik AI PandasAI versions up to 0.1.4 allows unauthenticated remote attackers to manipulate database operations through the pandasai-lancedb extension. Six functions (delete_question_and_answers, delete_docs, update_question_answer, update_docs, get_relevant_question_answers_by_id, get_relevant_docs_by_id) in lancedb.py are vulnerable to SQL injection attacks. Publicly available exploit code exists (CVSS 7.3, EPSS data not provided). The vendor has not responded to disclosure attempts.
Technical ContextAI
This vulnerability affects the pandasai-lancedb extension (cpe:2.3:a:sinaptik_ai:pandasai), specifically the LanceDB vector store integration used for semantic search and document retrieval in PandasAI's enterprise edition. The affected functions in extensions/ee/vectorstores/lancedb/pandasai_lancedb/lancedb.py perform database operations for managing question-answer pairs and document embeddings. The root cause is CWE-89 (SQL Injection), indicating insufficient input validation or parameterization when constructing SQL queries. Multiple CRUD operations (create, read, update, delete) across both Q&A and document management surfaces are vulnerable, suggesting systemic failure to sanitize user-controlled input before incorporating it into SQL statements. LanceDB is a vector database commonly used for AI/ML applications, and this integration layer appears to lack proper SQL parameterization or prepared statement usage.
RemediationAI
No vendor-released patch identified at time of analysis. The vendor was contacted early about this disclosure but has not responded or provided an official fix. Organizations should immediately implement the following mitigations: First, disable or remove the pandasai-lancedb extension if not operationally critical. Second, implement network-level access controls to restrict access to PandasAI installations to trusted IP ranges only, blocking unauthenticated internet access. Third, deploy web application firewall (WAF) rules to detect and block SQL injection attempts targeting the affected endpoints. Fourth, implement input validation and parameterized queries at the application layer if source code modification is feasible. Fifth, monitor database logs for suspicious query patterns targeting question-answer or document management functions. Sixth, consider replacing the vulnerable LanceDB integration with alternative vector store implementations until an official patch is released. For detailed vulnerability information and exploit proof-of-concept, refer to https://vuldb.com/?id.353883 and https://gist.github.com/YLChen-007/e33f76941234bed3824181ed252bd09f. Organizations should establish monitoring for vendor security announcements and upgrade immediately when patches become available.
GenerateSDFPipeline in synthetic_dataframe in PandasAI (aka pandas-ai) through 1.5.17 allows attackers to trigger the ge
An issue in Gaberiele Venturi pandasai v.0.8.0 and before allows a remote attacker to execute arbitrary code via a craft
An issue in pandas-ai v.0.9.1 and before allows a remote attacker to execute arbitrary code via the _is_jailbreak functi
Code injection in Sinaptik AI PandasAI versions up to 3.0.0 allows unauthenticated remote attackers to execute arbitrary
Path traversal in Sinaptik AI PandasAI up to version 3.0.0 allows remote unauthenticated attackers to read arbitrary fil
Same weakness CWE-89 – SQL Injection
View allShare
External POC / Exploit Code
Leaving vuln.today