// Security Advisory Triage

TOCTOU race in @anthropic-ai/claude-code allows file writes outside workspace

GitHub advisories — npm · published 2026-11-13 19:44 UTC
ACTION NEEDED PATCH_NOW SECURITY
WHAT CHANGEDGitHub published GHSA-5xjx-9qx6-jcrv on 9 November 2026, rating a time-of-check-time-of-use race condition in @anthropic-ai/claude-code as high severity. The vulnerability is tracked as CVE-2026-88968. Version 2.1.251 patches the issue.
WHY IT MATTERSThe race condition in the edit tool enables arbitrary file writes outside the intended workspace boundary. Services integrating Claude Code that process untrusted input or operate in multi-tenant environments risk unauthorized file system modification.
WHAT TO DOCheck whether @anthropic-ai/claude-code appears in your dependency tree. If present and below 2.1.251, update to 2.1.251 or later.
Model-reported confidence 85% (not measured accuracy) Primary source ↗ written by anthropic/claude-sonnet-4.5
Show the reasoning — raw diff, triage verdict, confidence
RAW DIFF
+ghsa_id: GHSA-5xjx-9qx6-jcrv
+summary: Claude Code: TOCTOU Race in Edit Tool Allows File Writes Outside the Workspace
+severity: high
+published_at: 2026-11-09T22:42:03Z
+cve_id: CVE-2026-88968
+vulnerabilities[].package.name: @anthropic-ai/claude-code
+vulnerabilities[].first_patched_version: 2.1.251
HOW IT WAS JUDGED
Triage signals below are model-reported assertions, preserved verbatim; they are not independent arithmetic verification.
triage verdictMATERIAL at 0.92, needed 0.5
triage reasoningHigh-severity vulnerability in Anthropic SDK affecting file system access control. Likely in dependency trees of AI-integrated services.
signalsAI provider SDK (@anthropic-ai/claude-code) · high severity · TOCTOU race condition enabling arbitrary file writes · new advisory with fixed version · likely in dependency tree of services using Claude Code integration
suppression—
triage modelanthropic/claude-haiku-4.5 ($0.00000)
writer modelanthropic/claude-sonnet-4.5 ($0.00000)