# CVE-2026-107291: Pydantic AI OpenTelemetry instrumentation content disclosure with include_content=False

Source: https://junglewise.ai/threats/cve-2026-107291-pydantic-ai-opentelemetry-instrumentation-exception-events-on  
Publisher: Junglewise Threat Intelligence, https://junglewise.ai/threats  
Published: 2026-10-08. Updated: 2026-10-10  
License: CC BY 4.0 with a link back

CVE-2026-107291 · Severity: medium · CVSS 4 · Published 2026-10-08

Pydantic AI is a framework for building AI agents with monitoring capabilities. When configured to exclude sensitive content from telemetry exports using the `include_content=False` setting, the OpenTelemetry instrumentation still leaks agent instructions, exception messages, error details, and output templates to observability backends. This bypasses the intended privacy control when traces are sent to third-party vendors or shared infrastructure.

Technologies: [Pydantic AI](https://junglewise.ai/threats/technologies/pydantic-pydantic-ai), [Pydantic AI Slim](https://junglewise.ai/threats/technologies/ai-slim), [pydantic-ai (PyPI)](https://junglewise.ai/threats/technologies/pydantic-ai), [pydantic-ai-slim (PyPI)](https://junglewise.ai/threats/technologies/pydantic-ai-slim). Vendors: [Pydantic](https://junglewise.ai/threats/vendors/pydantic), [PyPI](https://junglewise.ai/threats/vendors/pypi).

## Quick answers

**Is CVE-2026-107291 exploited in the wild?** No. Junglewise Threat Intelligence has no report of CVE-2026-107291 being exploited in the wild.

**How severe is CVE-2026-107291?** Medium severity, with a CVSS score of 4, published 8 October 2026.

**Which products and versions are affected?** Pydantic AI 0.3.4 to <1.107.6, 2.0.0b1 to <2.44.0; Pydantic AI Slim 0.3.4 to <1.107.6, 2.0.0b1 to <2.44.0; pydantic-ai (PyPI) ECOSYSTEM:introduced=0.3.4,fixed=1.107.6,introduced=2.0.0b1,fixed=2.44.0; pydantic-ai-slim (PyPI) ECOSYSTEM:introduced=0.3.4,fixed=1.107.6,introduced=2.0.0b1,fixed=2.44.0.

**Is there a fix for CVE-2026-107291?** Yes. pydantic-ai (PyPI): upgrade to 2.44.0 or later; pydantic-ai-slim (PyPI): upgrade to 2.44.0 or later. Until then: If you cannot upgrade, use an OpenTelemetry Collector processor to scrub exception.message, exception.stacktrace attributes, error status descriptions, and instruction parts and prompted_output_template within model_request_parameters on Pydantic AI spans. Alternatively, set include_model_request_parameters=False alongside include_content=False to remove both, though this loses other request parameter data. Junglewise finds every place you run it, drafts the upgrade and applies it once you approve.

**How is CVE-2026-107291 attacked?** It is attacked over the network.

**Am I affected by CVE-2026-107291?** Check whether you run Pydantic AI, Pydantic AI Slim, pydantic-ai (PyPI) or one of the others listed at an affected version. Junglewise does that check for you around the clock: connect your code, cloud and domains, and it matches new vulnerabilities like this one against what you actually run and tells you when one reaches you.

## How to fix

- pydantic-ai (PyPI): upgrade to 2.44.0 or later

  ```sh
  pip install "pydantic-ai>=2.44.0"
  ```

- pydantic-ai-slim (PyPI): upgrade to 2.44.0 or later

  ```sh
  pip install "pydantic-ai-slim>=2.44.0"
  ```


If you cannot upgrade, use an OpenTelemetry Collector processor to scrub exception.message, exception.stacktrace attributes, error status descriptions, and instruction parts and prompted_output_template within model_request_parameters on Pydantic AI spans. Alternatively, set include_model_request_parameters=False alongside include_content=False to remove both, though this loses other request parameter data.

## Technical details

The vulnerability is an information disclosure flaw (CWE-212, CWE-532) in Pydantic AI's OpenTelemetry instrumentation where exception events, error status descriptions, and model request parameters bypass the `include_content=False` redaction setting. Exception messages and stack traces on tool, agent run, model request, embedding, and image generation spans; OpenTelemetry error status descriptions; and serialized instructions and output templates in model_request_parameters are all exported despite the redaction flag. An attacker with access to exported telemetry can recover sensitive content (prompts, model responses, tool arguments, agent instructions) that should have been excluded.

## Affected products

- Pydantic pydantic-ai 0.3.4 to <1.107.6, 2.0.0b1 to <2.44.0
- Pydantic pydantic-ai-slim 0.3.4 to <1.107.6, 2.0.0b1 to <2.44.0

## Identifiers

- CVE-2026-107291
- GHSA-4x9p-g9wm-8q7f
- PYSEC-2026-4275
- PYSEC-2026-4284

## Timeline

- 2026-10-08: disclosed: Advisory published
- 2026-10-08: patched: Fixes released

## References

- [https://github.com/pydantic/pydantic-ai/security/advisories/GHSA-4x9p-g9wm-8q7f](https://github.com/pydantic/pydantic-ai/security/advisories/GHSA-4x9p-g9wm-8q7f)
- [https://github.com/pydantic/pydantic-ai/pull/8403](https://github.com/pydantic/pydantic-ai/pull/8403)

## Related threats

- [Pydantic AI concurrency slot leak in streamed requests](https://junglewise.ai/threats/cve-2026-107286-pydantic-ai-concurrency-limited-models-can-keep-their-slot-when)
- [Pydantic AI excessive resource consumption in HTML-to-Markdown conversion](https://junglewise.ai/threats/cve-2026-107287-pydantic-ai-excessive-resource-use-when-local-web-fetching)
- [Pydantic AI web chat UI Host header validation bypass](https://junglewise.ai/threats/cve-2026-107292-pydantic-ai-web-chat-ui-agent-to-web-clai-web-the-local-chat)
- [Pydantic AI missing content-type validation in chat endpoint](https://junglewise.ai/threats/cve-2026-107295-pydantic-ai-web-chat-ui-agent-to-web-clai-web-a-website-visited)
- [Pydantic AI OpenTelemetry instrumentation: retry prompt content is not redacted when `include_content=False`](https://junglewise.ai/threats/cve-2026-107293-pydantic-ai-opentelemetry-instrumentation-retry-prompt-content)