---
id: "CVE-2026-115726"
kind: "vulnerability"
aliases: ["CVE-2026-115726"]
title: "Aider: malicious committed config file runs shell commands on developer machines"
severity: "high"
cvss: 8.5
epss: 0.00249
exploited: false
status: "open"
category: "ai-tooling"
products: []
ecosystems: []
tags: ["ai-coding-agent", "ai-tooling", "aider", "developers"]
not_affected_if: ["not-using-aider", "only-trusted-repositories"]
published_at: "2026-10-23T12:18:13+00:00"
first_seen_at: "2026-10-23T21:57:32+00:00"
updated_at: "2026-10-25T12:38:51+00:00"
sources: [{"source": "nvd", "url": "https://nvd.nist.gov/vuln/detail/CVE-2026-115726", "first_seen_at": "2026-10-23T21:57:32+00:00"}, {"source": "epss", "url": "https://api.first.org/data/v1/epss?cve=CVE-2026-115726", "first_seen_at": "2026-10-25T12:38:51+00:00"}]
---

# Aider: malicious committed config file runs shell commands on developer machines

Aider, an AI coding assistant, auto-loads a .aider.conf.yml file from a repository without a trust prompt. An attacker who commits a malicious config can set notifications-command or editor values, which Aider runs through a shell after model replies or /editor use, executing commands with the victim's privileges.

_Summary, check and fix written by a model from the sources; labelled as such._

## How to check

- Run `aider --version` and compare against the fixed release in the NVD entry; the affected range is through `0.86.2`.
- Look for a `.aider.conf.yml` in cloned or pulled repositories and check for `notifications-command` or `editor` keys.

## Fix

- Upgrade Aider past `0.86.2`; the sources do not name the fixed version, so check the NVD entry or the Aider release notes.
- Until upgraded, delete or review any `.aider.conf.yml` in untrusted repositories before running Aider in them.

## Timeline

- 2026-10-23T21:57:32+00:00 nvd https://nvd.nist.gov/vuln/detail/CVE-2026-115726
- 2026-10-25T12:38:51+00:00 epss https://api.first.org/data/v1/epss?cve=CVE-2026-115726

## From the sources

Aider through 0.86.2 contains a command execution vulnerability that allows attackers to run shell commands by committing a malicious .aider.conf.yml that is auto-loaded without a trust prompt. Attackers can set notifications-command or editor values that are executed via subprocess with shell=True after model replies or /editor use, running commands with the victim's privileges.

---

_Free to read at a human pace. An API with keys and an MCP server are in preparation; early access: hello@infrastack.ai_