CVE-2026-114386
Pydantic AI Harness before 0.59.0 CodeMode Tool Call Limit Bypass via run_code
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.
Attack profile
How easy it is to exploit
- Attack vector
- Network
- Complexity
- Low
- Privileges needed
- None
- User interaction
- None
What an attacker can affect
- Confidentiality
- None
- Integrity
- Low
- Availability
- Low
Exploit likelihood
What each source reports
CVE List Score used
Pydantic AI Harness before 0.59.0 CodeMode Tool Call Limit Bypass via run_code
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.
- Record
CVE-2026-114386- CVSS 4.0
- 6.3 Medium
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 - CVSS 3.1
- 4.8 Medium
CVSS:3.1/AV:N/AC:H/PR:N/UI:N/S:U/C:N/I:L/A:L - Weakness
- CWE-770
- Updated
- 2026-11-02
EUVDEUVD-2026-107233
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.
- Record
EUVD-2026-107233- CVSS 4.0
- 6.3 Medium
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 - Updated
- 2026-11-02
GitHub AdvisoriesGHSA-3589-pxp7-3gj8
Pydantic AI Harness tool-call limit bypass in nested CodeMode calls
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.
- Record
GHSA-3589-pxp7-3gj8- CVSS 4.0
- 6.3 Medium
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/E:X/CR:X/IR:X/AR:X/MAV:X/MAC:X/MAT:X/MPR:X/MUI:X/MVC:X/MVI:X/MVA:X/MSC:X/MSI:X/MSA:X/S:X/AU:X/R:X/V:X/RE:X/U:X - CVSS 3.1
- 4.8 Medium
CVSS:3.1/AV:N/AC:H/PR:N/UI:N/S:U/C:N/I:L/A:L - Weakness
- CWE-770
- Updated
- 2026-11-02
NVD
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.
- Record
CVE-2026-114386- CVSS 4.0
- 6.3 Medium
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/E:X/CR:X/IR:X/AR:X/MAV:X/MAC:X/MAT:X/MPR:X/MUI:X/MVC:X/MVI:X/MVA:X/MSC:X/MSI:X/MSA:X/S:X/AU:X/R:X/V:X/RE:X/U:X - CVSS 3.1
- 4.8 Medium
CVSS:3.1/AV:N/AC:H/PR:N/UI:N/S:U/C:N/I:L/A:L - Weakness
- CWE-770
- Updated
- 2026-11-02
OSV
Pydantic AI Harness before 0.59.0 CodeMode Tool Call Limit Bypass via run_code
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.
- Record
CVE-2026-114386- CVSS 4.0
- 6.3 Medium
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 - Weakness
- CWE-770
- Updated
- 2026-11-02
OSVGHSA-3589-pxp7-3gj8
Pydantic AI Harness tool-call limit bypass in nested CodeMode calls
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.
- Record
GHSA-3589-pxp7-3gj8- CVSS 4.0
- 6.3 Medium
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/E:X/CR:X/IR:X/AR:X/MAV:X/MAC:X/MAT:X/MPR:X/MUI:X/MVC:X/MVI:X/MVA:X/MSC:X/MSI:X/MSA:X/S:X/AU:X/R:X/V:X/RE:X/U:X - CVSS 3.1
- 4.8 Medium
CVSS:3.1/AV:N/AC:H/PR:N/UI:N/S:U/C:N/I:L/A:L - Weakness
- CWE-770
- Updated
- 2026-11-02
Affected products
Affected packages
| Package | Affected | Fixed |
|---|---|---|
pydantic-ai-harnessPyPI |
before 0.59.0 | 0.59.0 |
pydantic-ai-harnessPyPI |
before 0.59.0, plus listed versions 0.1.10.1.20.10.00.11.00.12.0 Show 43 moreShow less0.13.00.14.00.15.00.16.00.17.00.18.00.18.10.2.00.2.10.20.00.21.00.22.00.23.00.24.00.25.00.26.00.27.00.28.00.28.10.29.00.29.10.3.00.30.00.31.00.32.00.33.00.34.00.35.00.36.00.4.00.5.00.52.00.53.00.54.00.55.00.56.00.57.00.58.00.6.00.7.00.7.10.8.00.9.0 |
0.59.0 |
History
| Recorded | Changed | Sources |
|---|---|---|
| first indexed | CVE List, EUVD, GitHub Advisories, NVD, OSV |
Dates are when Every CVE recorded each change. Upstream dates may differ.
In short
CVE-2026-114386 is a medium vulnerability (CVSS 6.3) in Pydantic Pydantic-ai-harness, published October 21, 2026. No source lists public exploit code. EPSS puts the chance of exploitation in the next 30 days at 0.35%, the 28th percentile. Fixed in pydantic-ai-harness 0.59.0.
References
github.com/pydantic/pydantic-ai/pull/10362
pypi.org/project/pydantic-ai-harness/0.59.0
vulncheck.com/advisories/pydantic-ai-harness-before-0.59.0-codemode-tool-call-limit-bypass-via-run-code