CVE-2026-57456: Vim Python Omni-Completion Code Execution Vulnerability

Technical summary

The vulnerability exists in Vim's Python 3 omni-completion module (runtime/autoload/python3complete.vim and legacy pythoncomplete.vim), which uses exec() to evaluate reconstructed Python definitions for completion population. When reconstructing function and class scopes, docstrings are inserted verbatim between triple quotes without any escaping or sanitization. An attacker can inject a docstring containing triple quotes followed by arbitrary Python code to escape the quoted literal and achieve code execution within the Python interpreter context that Vim spawns. The flaw is a classic code injection issue (CWE-94: Improper Control of Generation of Code) where untrusted input from the buffer is reconstructed into executable code without proper escaping.

Business impact

Any user opening a Vim buffer containing a hostile Python file and triggering omni-completion will execute the attacker's code with the privileges of the Vim process. For developers and system administrators who rely on Vim, this creates a supply-chain or file-handling risk: cloning a repository, opening a Python file, and pressing Ctrl+X Ctrl+O (or equivalent completion keystroke) could compromise the user's session. The impact is heightened in shared editing environments or when reviewing untrusted code. Malicious code runs in the user's security context, potentially allowing data theft, credential harvesting, or lateral movement.

Affected systems

Vim versions prior to 9.2.0699 are affected. Any installation with Python 3 omni-completion enabled and the pythoncomplete or python3complete modules active can be exploited. The vulnerability is not specific to a particular operating system; it affects Vim on Linux, macOS, Windows, and other platforms where Vim and Python 3 are available. Users with Vim compiled without Python support are not affected.

Exploitability

Exploitability is straightforward. The attack requires local code execution (AV:L) and user interaction (UI:R)—specifically, the user must open a hostile buffer and trigger omni-completion. No authentication or special privileges are needed before the vulnerable code path is reached. Once triggered, code execution is immediate and reliable. The barrier to exploitation is low: an attacker commits a malicious .py file to a public repository, and any developer who clones it and opens the file in Vim faces risk. Tools that automatically trigger completion (e.g., vim-lsp or similar plugins) could lower the user-interaction requirement further.

Remediation

Update Vim to version 9.2.0699 or later. This release includes fixes to properly escape docstrings during function and class reconstruction, preventing the triple-quote breakout. Users should verify their installed version with :version in Vim and consult the official Vim release notes to confirm the patch is included. For distributions that package Vim, check with your package manager for availability of the patched version. In the interim, disabling Python omni-completion (by not loading python3complete.vim or limiting its scope) reduces risk, though this is a workaround rather than a fix.

Patch guidance

The fix is included in Vim 9.2.0699. Verify your version by running :version inside Vim and checking the included patch number. Most Linux distributions should release updates within days of the Vim upstream release. macOS users can update via Homebrew (brew upgrade vim) or MacPorts. Windows users can download the latest installer from vim.org. After patching, restart Vim and re-test omni-completion to ensure it functions as expected. If you maintain a custom Vim build, apply the patch from the official Vim repository and recompile.

Detection guidance

Monitor for unexpected Python subprocess spawning from Vim, unusual process executions triggered by omni-completion (Ctrl+X Ctrl+O), or signs of code execution during editor sessions. On systems with Python audit logging enabled, look for exec() or eval() calls originating from Vim's Python context with suspicious or obfuscated payloads. Review Vim logs and terminal history for evidence of completion being triggered on untrusted files. Network-based detection is limited since code execution is local; focus on endpoint logs, process execution monitoring (EDR/XDR tools), and file integrity checks for unexpected modifications during or after Vim sessions.

Why prioritize this

This vulnerability scores 7.8 (HIGH) due to the combination of high-impact code execution (C:H, I:H, A:H), low attack complexity, and minimal barriers to exploitation. While the attack requires user interaction, that interaction—opening a file and using a standard editor feature—is routine for developers. The attack surface is broad (any Vim user with Python completion enabled), and the consequences are severe (full code execution). Organizations with developer populations using Vim should treat this as medium-to-high priority for patching, particularly if developers handle untrusted code or work in collaborative environments.

Risk score, explained

The CVSS 3.1 score of 7.8 reflects a HIGH severity rating. Attack Vector is Local (AV:L) because the attacker must deliver a malicious file to the target system; Attack Complexity is Low (AC:L) because exploiting the code path requires no special conditions beyond opening a file and triggering completion. Privileges Required are None (PR:N), and User Interaction is Required (UI:R) because the developer must manually trigger omni-completion. The scope is Unchanged (S:U). The impact across Confidentiality, Integrity, and Availability is High (C:H, I:H, A:H) because code execution grants full control over the Vim process and any child processes, enabling data exfiltration, system compromise, and denial of service.

Frequently asked questions