CVE-2026-114386 json
Pydantic AI Harness before 0.59.0 CodeMode Tool Call Limit Bypass via run_code
published 2026-10-21 19 d ago modified 2026-10-23 16 d ago also GHSA-3589-pxp7-3gj8 , EUVD-2026-107233
- cvss
- 4.8
- version 3.1, CNA
- cisa kev
- Not listed
- not in CISA's catalog
- epss
- 0.3%
- EPSS 0.3% chance of exploitation in 30 days, percentile 28%
- public exploits
- 0
- none known
Summary
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.
Patch / fix
- pypi pydantic-ai-harness 0.59.0
- pydantic pydantic-ai-harness 0.59.0
CISA SSVC
- Exploitation
- none
- Automatable
- no
- Technical impact
- partial
- Assessed
- 18 d ago
Exploit prediction
0.35% chance of exploitation in the next 30 days, higher than 28% of all scored CVEs (FIRST EPSS, 2026-11-08).
CVSS scores
| assigned by | version | score | vector |
|---|---|---|---|
| CNA | 4.0 | 6.3 | 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 |
| CNA | 3.1 | 4.8 | CVSS:3.1/AV:N/AC:H/PR:N/UI:N/S:U/C:N/I:L/A:L |
| CNA | 4.0 | 6.3 | 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 |
| GitHub | 3.1 | 4.8 | CVSS:3.1/AV:N/AC:H/PR:N/UI:N/S:U/C:N/I:L/A:L |
Affected packages
| ecosystem | package | affected | fixed in |
|---|---|---|---|
| pypi | pydantic-ai-harness | <0.59.0 |
0.59.0 |
Affected products
- pydantic pydantic-ai-harness
Classification
- Weakness
-
- CWE-770 Allocation of Resources Without Limits or Throttling
References
-
github.com
/pydantic/pydantic-ai/pull/10362
issue-tracking, patch, web
-
github.com
/pydantic/pydantic-ai/commit/eb27b2b89fe76ae23ee9b6fc9eb4b227a4967949
patch, web
-
github.com
/pydantic/pydantic-ai/releases/tag/v2.59.0
release-notes, web
-
github.com
/pydantic/pydantic-ai/blob/v2.58.0/src/pydantic_ai_harness/pydantic_ai_harness/code_mode/_toolset.py#L642-L680
technical-description, web
-
pypi.org
/project/pydantic-ai-harness/0.59.0/
product
-
vulncheck.com
/advisories/pydantic-ai-harness-before-0.59.0-codemode-tool-call-limit-bypass-via-run-code
third-party-advisory, web
-
nvd.nist.gov
/vuln/detail/CVE-2026-114386
advisory
-
github.com
/pydantic/pydantic-ai
package
-
pypi.org
/project/pydantic-ai-harness/0.59.0
web
Sources
- CVE List CVE-2026-114386
- GitHub Advisory GHSA-3589-pxp7-3gj8
- NVD CVE-2026-114386
Fencer Vulns keeps one entry per vulnerability: the CVE id wins, then GHSA, and records that alias each other are merged, except that two CVEs always keep separate entries. The same data is available as JSON.