CVE-2026-27001

The record

OpenClaw is a personal AI assistant. Prior to version 2026.2.15, OpenClaw embedded the current working directory (workspace path) into the agent system prompt without sanitization. If an attacker can cause OpenClaw to run inside a directory whose name contains control/format characters (for example newlines or Unicode bidi/zero-width markers), those characters could break the prompt structure and inject attacker-controlled instructions. Starting in version 2026.2.15, the workspace path is sanitized before it is embedded into any LLM prompt output, stripping Unicode control/format characters and explicit line/paragraph separators. Workspace path resolution also applies the same sanitization as defense-in-depth.

Affected packages

Related vulnerabilities

• GHSA-2qj5-gwg2-xwc4 OpenClaw: Unsanitized CWD path injection into LLM prompts Same vulnerability
• GHSA-4hpg-mp64-x7xq OpenClaw: Internal/webchat command auth could inherit ownerAllowFrom wildcard state Also affects openclaw
• GHSA-8mg9-j9cf-54cj OpenClaw: Empty-scope device re-pairing could confuse caller scope containment Also affects openclaw
• GHSA-rx78-29qr-5hq8 OpenClaw: Workspace-derived service PATH could influence trash command selection Also affects openclaw
• GHSA-wc84-j36w-pw4x OpenClaw: Workspace .env STATE_DIRECTORY could influence bundled runtime dependency roots Also affects openclaw
• GHSA-cw4q-gqg5-g38h OpenClaw: Discord allowFrom could bind to mutable display names Also affects openclaw
• GHSA-mpc8-jxjh-qpgh OpenClaw: Focus command could miss controlScope enforcement Also affects openclaw
• GHSA-24vr-rprv-67rf OpenClaw: Workspace .env npm_execpath could influence bundled runtime dependency install Also affects openclaw