Browse the dataset
- 2,412 vulnerabilities
- 2,412 vulnerable / patched code pairs
- 930 exploit payloads
- 10 categories
- newest CVE
Active filters
| CVE | Summary | Category | CWE | Scope | CVSS | Published |
|---|---|---|---|---|---|---|
| CVE-2026-109440 | Indirect prompt injection lets attackers run shell/Python, exfiltrate vault secrets. | Input Validation and Sanitization Command Injection | CWE-94 | Remote | 8.7 | |
| CVE-2026-51857 | CodeExecutionToolkit runs modelโgenerated Python via SubprocessInterpreter without approval. | Input Validation and Sanitization Command Injection | CWE-94 | Remote | Not scored | |
| CVE-2026-51856 | Remote code execution via RealtimeAgent execute_python_code tool exposure. | Authentication, Authorization, and Session Management Privilege Escalation | CWE-94 | Remote | Not scored | |
| CVE-2026-53710 | Unsafe raw getattr in python_sandbox_server lets attackers exec OS commands via crafted dunder names. | Design Defects Poorly Designed Access Controls | CWE-94 | Remote | 10.0 | |
| CVE-2026-85694 | Remote code execution via untrusted LLM output in LaVagueโฏ0.2.35โs extractor. | Input Validation and Sanitization Command Injection | CWE-94 | Remote | 9.2 | |
| CVE-2026-51974 | Eval injection in get_list (meta_parser.py) via crafted EXIF styles payload. | Input Validation and Sanitization Command Injection | CWE-94 | Remote | 8.8 | |
| CVE-2026-82278 | Remote code execution via unauthenticated workflow run_once endpoint (exec() abuse). | Input Validation and Sanitization Command Injection | CWE-94 | Remote | 8.7 | |
| CVE-2026-37003 | Remote code execution via unsanitized LLM prompts in Agno (โค2.5.8) components. | Input Validation and Sanitization Command Injection | CWE-94 | Remote | 9.8 | |
| CVE-2026-81096 | Unauthenticated sandbox escape via dunder attribute injection in Python executor. | Authentication, Authorization, and Session Management Insecure Authentication Mechanisms | CWE-94 | Remote | 9.3 | |
| CVE-2026-58474 | Code injection via crafted GGUF filename in whichllmย 0.5.16 CLI commands. | Input Validation and Sanitization Command Injection | CWE-94 | Remote | 8.6 | |
| CVE-2026-57170 | Serverโside template injection via nonโsandboxed Jinja2 include tags in Trestle. | Input Validation and Sanitization Insecure Parsing or Deserialization | CWE-94 | Remote | 7.8 | |
| CVE-2026-54757 | Serverโside template injection in Trestle Jinja2 tags enables remote code execution. | Input Validation and Sanitization Command Injection | CWE-94 | Remote | 7.8 | |
| CVE-2026-78379 | LLM prompt injection in python_repl lets attackers run arbitrary Python code before 0.8.5. | Input Validation and Sanitization Command Injection | CWE-94 | Remote | 9.2 | |
| CVE-2026-55585 | RCE via attacker-controlled math expressions passed to SymPy parse_expr. | Input Validation and Sanitization Insecure Parsing or Deserialization | CWE-94 | Remote | 8.8 | |
| CVE-2026-55546 | Arbitrary code execution via unfiltered SymPy expression parsing in verify_math_expression(). | Input Validation and Sanitization Insecure Parsing or Deserialization | CWE-94 | Remote | 9.8 | |
| CVE-2026-76841 | Xinference enables remote code execution via HuggingFace models by default. | Configuration Issues Remote File Inclusion (RFI) | CWE-94 | Remote | 8.7 | |
| CVE-2026-68508 | Arbitrary code execution via untrusted Hydra instantiate target values before v1.3.4. | Input Validation and Sanitization Insecure Parsing or Deserialization | CWE-94 | Remote | 7.8 | |
| CVE-2026-62675 | Arbitrary command execution via malicious agent bundle path injection in Omnigent โค0.3.0. | Input Validation and Sanitization Command Injection | CWE-94 | Remote | 8.8 | |
| CVE-2026-18287 | Remote code execution via unvalidated user string in load_time_series_segmentation_benchmark. | Input Validation and Sanitization Command Injection | CWE-94 | Remote | 7.8 | |
| CVE-2026-18286 | Remote code execution via unvalidated string in load_human_activity_segmentation_datasets. | Input Validation and Sanitization Command Injection | CWE-94 | Remote | 7.8 | |
| CVE-2026-45272 | Admin can inject Python code via unchecked SOCIAL_AUTH keys, leading to RCE. | Input Validation and Sanitization Command Injection | CWE-94 | Remote | 9.4 | |
| CVE-2026-75858 | Remote code execution via rlm_eval autoโapproval bypass in CodeWhale 0.8.41โ0.8.63. | Design Defects Poorly Designed Access Controls | CWE-94 | Remote | 8.5 | |
| CVE-2026-0765 | Plugins extension runs userโprovided Python with granted permissionโintended behavior. | Design Defects Poorly Designed Access Controls | CWE-94 | Remote | Not scored | |
| CVE-2026-59894 | sqlparse allows code injection via unescaped backslashes in formatted output. | Input Validation and Sanitization Command Injection | CWE-94 | Remote | 6.2 | |
| CVE-2026-73678 | MindsDB unauth RCE via crafted prompts to the agent's scratchpad tool. | Authentication, Authorization, and Session Management Command Injection | CWE-94 | Remote | 10.0 |
Download the dataset
Every visible entry with all of its fields: classification, descriptions, the vulnerable and patched code pair, and the exploit payload where there is one. Current files: 2,411 entries, built .
- CSV For spreadsheets, pandas and R. One row per entry; code samples are quoted fields. 9.0 MB
- JSON Records plus licence, citation and field descriptions in one file. 10.0 MB
SHA-256 checksums
00100ae17adddda8f786a502580a64470cdaa8e95d6d33eded83ffdbdfc63386 vaitp-dataset-2026-11-21.csv
77624d5fe7cf1f5e7dde13d81a8186fd8fc520a506b7dedf831e3ffc6a232722 vaitp-dataset-2026-11-21.json
Licence and citation
The VAITP dataset is published under the Creative Commons Attribution 4.0 International (CC BY 4.0) licence: you may share and adapt it for any purpose, with attribution.
Bogaerts, F., Ivaki, N., & Fonseca, J. (2026). VAITP Python Vulnerability Dataset [Data set]. https://netpack.pt/vaitp/dataset/
@misc{vaitp_dataset_2026,
title = {{VAITP} Python Vulnerability Dataset},
author = {Bogaerts, Fr\'ed\'eric and Ivaki, Naghmeh and Fonseca, Jos\'e},
year = {2026},
howpublished = {\url{https://netpack.pt/vaitp/dataset/}},
note = {Licensed under CC BY 4.0}
}
What each field means
| Field | Meaning |
|---|---|
id | Internal VAITP record number. |
cve | CVE identifier. |
cve_url | Link to the CVE record. |
cwe | CWE weakness as stored, e.g. "CWE-79" or "CWE-94 (inferred)" when assigned by the labelling model rather than NVD. |
cwe_number | Numeric CWE identifier, when there is one. |
cwe_url | Link to the CWE definition at MITRE. |
cvss_score | CVSS base score from NVD (newest version available). Empty when NVD has not scored it. |
cvss_rating | Qualitative rating of the score: low, medium, high or critical. |
published | CVE publication date (YYYY-MM-DD). |
owasp | OWASP Top 10 category. |
odc | Orthogonal Defect Classification defect type. |
cdc | Code defect classification (VAITP taxonomy). |
category | Vulnerability category (VAITP taxonomy). |
subcategory | Vulnerability subcategory (VAITP taxonomy). |
scope | Accessibility scope: whether the flaw is reachable remotely or only locally. |
impact | Main impact of a successful exploit. |
summary | One-line description. |
description | Full CVE description. |
solution | How the vulnerability is fixed or mitigated. |
affected_component | Library, package or component the change is in. |
fixed_by_upgrading | Whether upgrading the affected component fixes it. |
vulnerable_code | Python sample containing the vulnerability. |
patched_code | The same sample with the vulnerability fixed. |
payload_needed | Whether exploiting the sample needs a crafted input. |
payload | Proof-of-concept input that exploits the vulnerable sample, when available. |
payload_status | available, none, or model_refused (the labelling model declined to write one). |
references | Further references. |
permalink | This entry's page on the VAITP website. |
Use the API instead
The same filters work on the read-only JSON API, so you can script against it:
https://netpack.pt/vaitp/wp-json/vaitp/v1/vulnerabilities?category=Cryptographic&min_score=7&per_page=100
https://netpack.pt/vaitp/wp-json/vaitp/v1/vulnerabilities/CVE-2024-1234
https://netpack.pt/vaitp/wp-json/vaitp/v1/stats
https://netpack.pt/vaitp/wp-json/vaitp/v1/export?format=csv&q=flask
Introducing the "VAITP dataset": a specialized repository of Python vulnerabilities and patches, meticulously compiled for the use of the security research community. As Python's prominence grows, understanding and addressing potential security vulnerabilities become crucial. Crafted by and for the cybersecurity community, this dataset offers a valuable resource for researchers, analysts, and developers to analyze and mitigate the security risks associated with Python. Through the comprehensive exploration of vulnerabilities and corresponding patches, the VAITP dataset fosters a safer and more resilient Python ecosystem, encouraging collaborative advancements in programming security.
The supreme art of war is to subdue the enemy without fighting.
Sun Tzu – “The Art of War”
:: Shaping the future through research and ingenuity ::
