CVE-2026-114386
Published:October 21, 2026
Updated:October 23, 2026
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.
Affected Packages
pydantic-ai-harness (CONDA):
Affected version(s) >=0.1.1 <0.59.0Fix Suggestion:
Update to version 0.59.0https://github.com/pydantic/pydantic-ai.git (GITHUB):
Affected version(s) >=v2.52.0 <v2.59.0Fix Suggestion:
Update to version v2.59.0pydantic-ai-harness (PYTHON):
Affected version(s) >=0.1.1 <0.59.0Fix Suggestion:
Update to version 0.59.0Related Resources (9)
Do you need more information?
Contact UsCVSS v4
Base Score:
6.3
Attack Vector
NETWORK
Attack Complexity
LOW
Attack Requirements
PRESENT
Privileges Required
NONE
User Interaction
NONE
Vulnerable System Confidentiality
NONE
Vulnerable System Integrity
LOW
Vulnerable System Availability
LOW
Subsequent System Confidentiality
NONE
Subsequent System Integrity
NONE
Subsequent System Availability
NONE
CVSS v3
Base Score:
4.8
Attack Vector
NETWORK
Attack Complexity
HIGH
Privileges Required
NONE
User Interaction
NONE
Scope
UNCHANGED
Confidentiality
NONE
Integrity
LOW
Availability
LOW
Weakness Type (CWE)
Allocation of Resources Without Limits or Throttling
EPSS
Base Score:
0.35