Pydantic AI OpenTelemetry instrumentation: retry prompt content is not redacted when `include_content=False`

Summary

Pydantic AI is a Python agent framework for building applications and workflows with Generative AI. From 0.3.4 until 1.107.4 and 2.27.1, OpenTelemetry instrumentation configured with InstrumentationSettings(include_content=False) can export retry prompts outside tool calls in gen_ai.input.messages and pydantic_ai.all_messages. Agents using NativeOutput, PromptedOutput, or output validators on text output can therefore disclose validation feedback, including invalid model values quoted by that feedback, to readers of the telemetry backend. Tool-call retries and deployments that do not use include_content=False are not affected by this specific path. This issue is fixed in versions 1.107.4 and 2.27.1.

Timeline

1. Oct 8, 2026, 4:57 PM UTCCVE published
2. Oct 8, 2026, 5:14 PM UTCCISA SSVC assessment

   Exploitation: none · Automatable: no · Technical impact: partial

3. Oct 8, 2026, 5:14 PM UTCRecord last updated

Affected versions

• pydantic pydantic-ai

  From 0.3.4 · fixed in 1.107.4; From 2.0.0b1 · fixed in 2.27.1

• pydantic pydantic-ai-slim

  From 0.3.4 · fixed in 1.107.4; From 2.0.0b1 · fixed in 2.27.1

• pydantic-aiPyPI

  From 0.3.4 · fixed in 1.107.4

• pydantic-aiPyPI

  From 2.0.0b1 · fixed in 2.27.1

• pydantic-ai-slimPyPI

  From 0.3.4 · fixed in 1.107.4

• pydantic-ai-slimPyPI

  From 2.0.0b1 · fixed in 2.27.1

• pydantic-aiPyPI

  From 0.3.4 · fixed in 1.107.4; From 2.0.0b1 · fixed in 2.27.1

• pydantic-ai-slimPyPI

  From 0.3.4 · fixed in 1.107.4; From 2.0.0b1 · fixed in 2.27.1

References 4