VYPR
Critical severity9.1GHSA Advisory· Published Jun 23, 2026· Updated Jun 29, 2026

CVE-2026-54316

CVE-2026-54316

Description

Claude Code is an agentic coding tool. From 0.2.54 until 2.1.163, because the hostname huggingface.co was pre-approved as a bare hostname for the WebFetch tool, any path on that domain—including attacker-controlled model repositories—was auto-approved without a permission prompt or being subject to --allowedTools restrictions. An attacker able to inject untrusted content into a Claude Code context could direct it to issue WebFetch requests against attacker-controlled repository files (e.g. /resolve/main/config.json), which HuggingFace counts as downloads server-side, creating a covert out-of-band channel for encoding and exfiltrating data Claude can access such as files, environment variables, or command output. Reliably exploiting this required the ability to add untrusted content into a Claude Code context window. This vulnerability is fixed in 2.1.163.

AI Insight

LLM-synthesized narrative grounded in this CVE's description and references.

Affected packages

Versions sourced from the GitHub Security Advisory.

PackageAffected versionsPatched versions
@anthropic-ai/claude-codenpm
>= 0.2.54, < 2.1.1632.1.163

Affected products

3
  • cpe:2.3:a:anthropic:claude_code:*:*:*:*:*:node.js:*:*+ 1 more
    • cpe:2.3:a:anthropic:claude_code:*:*:*:*:*:node.js:*:*range: >=0.2.54,<2.1.163
    • (no CPE)range: >= 0.2.54, < 2.1.163
  • ghsa-coords
    Range: >= 0.2.54, < 2.1.163

Patches

Vulnerability mechanics

References

3

News mentions

3