CVE-2026-114386 MEDIUM

CVE-2026-114386: Pydantic AI Harness before 0.59.0 CodeMode Tool Call Limit Bypass via run_code

Vendor Pydantic

Product pydantic-ai-harness

Weakness CWE-770 · Uncontrolled resource consumption

Published October 21, 2026

Last update October 21, 2026

View on NVD All CVEs

CVSS base score

6.3/10

Attack vector Network

Attack complexity Low

Privileges required None

User interaction None

Confidentiality —

Integrity —

CVSS vector

CVSS:4.0/AV:N/AC:L/AT:P/PR:N/UI:N/VC:N/VI:L/VA:L/SC:N/SI:N/SA:N

What the vulnerability does

01Description

Pydantic AI Harness before 0.59.0 fails to enforce UsageLimits.tool_calls_limit on nested tool calls dispatched from run_code in the CodeMode toolset. Users or prompt-injected content can make the sandboxed program invoke tools via run_nested_call() without check_before_tool_call, exceeding operator limits on costly or side-effecting tools.

Key dates

02Disclosure timeline

October 21, 2026 CVE published

October 21, 2026 Record updated

External resources

03References

NVD — National Vulnerability Database https://nvd.nist.gov/vuln/detail/CVE-2026-114386 CWE — Common Weakness Enumeration https://cwe.mitre.org/data/definitions/770.html AskarLabs — CWE-770 write-up Allocation of Resources Without Limits or Throttling

Related vulnerabilities

CVE-2026-114386: Pydantic AI Harness before 0.59.0 CodeMode Tool Call Limit Bypass via run_code

01Description

02Disclosure timeline

03References

04Related CVE