A misinterpretation of the workspace trust flag in vscode‑spell‑checker allows an attacker to place a malicious .cspell.config.js in a workspace that is then loaded automatically when the workspace is opened. The configuration value cSpell.trustedWorkspace is treated as an authoritative trust flag by default, and any truthy value is coerced to true. This bypass lets untrusted workspaces run attacker‑controlled JavaScript/TypeScript code with the user’s privileges, providing an arbitrary code execution vector. The underlying weakness stems from improper trust validation and incorrect configuration handling (CWE‑276, CWE‑807, CWE‑829).

The affected product is the Streetsidesoftware vscode‑spell‑checker extension for Visual Studio Code. All releases prior to v4.5.4 are vulnerable, including v4.5.0 to v4.5.3 and earlier minor versions. The vulnerability is limited to systems where the extension is installed and a malicious workspace is opened.

The CVSS score is 7.8, indicating high severity. The EPSS score is below 1%, reflecting a low probability of exploitation in the wild. It is not listed in CISA’s KEV catalog, but due to the local nature of the attack, an attacker can trigger the flaw by creating or modifying a workspace. The primary attack vector is a maliciously crafted workspace opened by an ordinary user; no remote network access is required.