← EVENT LIBRARY / Developer Tools

Pydantic AI closes web-fetch policy and telemetry leaks

Pydantic AI 2.44.0 fixes four security issues in web fetching and telemetry. Two bypass network or domain policy through IPv6 zone identifiers or hostname normalization. One attacker-controlled page can trigger quadratic processing and block the event loop. Another can export instructions and output templates even when telemetry content is disabled. Version 1.107.6 backports the fixes.

Imported from the existing research archive. Migration preserves the recorded judgment and is not a fresh verification.

What this means for you

For production agents that fetch arbitrary URLs, these are runtime-boundary flaws rather than cosmetic patches. Upgrade both current and v1 deployments, then test private-address and normalized-hostname cases in local policy. Teams relying on content-redacted telemetry should also review already-exported traces.

Evidence & related material

Background · Historical record, not rechecked

Pydantic AI 2.44.0 fixes four security issues in web fetching and telemetry. Two bypass network or domain policy through IPv6 zone identifiers or hostname normalization. One attacker-controlled page can trigger quadratic processing and block the event loop. Another can export instructions and output templates even when telemetry content is disabled. Version 1.107.6 backports the fixes.

Subsequent changes

Update · 2026-09-26

Pydantic AI 2.51.0 adds GPT-Live support and realtime context usage while reducing agent graph overhead.

Technical details & scores
Fixed Versions 2.44.0 and v1 backport 1.107.6
Issues IPv6 zone SSRF bypass; normalized-host blocklist bypass; quadratic web_fetch; telemetry content leak
Severity two moderate, two low
Action upgrade and retest network policy plus telemetry redaction
Technical impact 3
Engineering value 5
Adoption signal 4
Maturity 4
Verification cost 1
Risk 4