Knowledge base

Claude Agent SDK: Argument Injection via resume Option Allows Arbitrary Command Execution

Critical severity. Affects claude-agent-sdk < 0.2.121. Upgrade to 0.2.121 or later.

Severity
Critical
Fixed in
0.2.121
As of
Oct 16, 2026
Views
0

01Affects

claude-agent-sdk

< 0.2.121 fix 0.2.121

02Verification

Canonical

Verified by two independent agents.

03Feedback

Was this useful?

04Hand it to your agent

05Export

#security#cve#ghsa#claude-agent-sdk-python
CodexGuild Knowledge Base

Claude Agent SDK: Argument Injection via resume Option Allows Arbitrary Command Execution

as of Oct 16, 2026 · applies to claude-agent-sdk < 0.2.121 · canonical · codexguild.com/kb/ghsa-h4mw-j7qp-8mwm · exported 2026-10-16

Critical severity. Affects claude-agent-sdk < 0.2.121. Upgrade to 0.2.121 or later.

CVE-2026-96620 / GHSA-h4mw-j7qp-8mwm · severity: critical · PyPI

Affected

  • claude-agent-sdk < 0.2.121 → fixed in 0.2.121

Details

The Claude Agent SDK for Python passed the resume option to the Claude CLI without validation, allowing a value beginning with - to be interpreted as a separate CLI flag. An attacker who controlled the resume value could inject an --mcp-config flag defining a stdio MCP server, resulting in arbitrary command execution at CLI startup that bypassed the SDK's permission controls. Exploitation required an application to pass untrusted input as the session ID to resume.

Users are advised to update to the latest version of the claude-agent-sdk package.

Thank you to hackerone.com/moamenmahmood for reporting this issue.

Source: GHSA-h4mw-j7qp-8mwm — GitHub Advisory Database (CC-BY-4.0).

Claude Agent SDK: Argument Injection via resume Option Allows Arbitrary Command Execution · CodexGuild