How to Fix CVE-2026-25640: Path Traversal in pydantic-ai
Related fixes
Other vulnerabilities in the same area that are worth patching alongside this one:
- How to Fix CVE-2026-2859: CWE-204: Observable Response Discrepancy in Checkmk — CWE-204: Observable Response Discrepancy in Checkmk
- How to Fix CVE-2026-29974: Stack-based Buffer Overflow in the affected product — Stack-based Buffer Overflow in the affected product
- How to Fix CVE-2026-1945: Cross-site scripting in WPBookit , Cross-site scripting in WPBookit
- How to Fix CVE-2026-4880: privilege management flaw in Barcode Scanner (+Mobile App) – Inventory manager, Order fulfillment system, POS (Point of Sale) , privilege management flaw in Barcode Scanner (+Mobile App) – Inventory manager, Order fulfillment system, POS (Point of Sale)
- How to Fix CVE-2026-3839: CWE-22: Improper Limitation of a Pathname to a Restricted Directory ('Path Traversal') , CWE-22: Improper Limitation of a Pathname to a Restricted Directory ('Path Traversal')
*By Sai Kiran Pandrala*
| Severity | CVSS 7.1 - High |
|---|---|
| Actively exploited? | Not currently listed in CISA KEV |
| Affected | >= 1.34.0, < 1.51.0 |
| Fixed in | 1.51.0. |
| Type (CWE) | CWE-22: Improper Limitation of a Pathname to a Restricted Directory ('Path Traversal') |
What is CVE-2026-25640?
CVE-2026-25640 is a path traversal flaw in pydantic-ai. The product fails to canonicalize or restrict file paths supplied by a remote caller, so .. sequences or absolute paths reach restricted parts of the filesystem. Vendor description: Pydantic AI is a Python agent framework for building applications and workflows with Generative AI. From 1.34.0 to before 1.51.0, a path traversal vulnerability in the Pydantic AI web UI allows an attacker to serve arbitrary JavaScript in the context of the application by crafting a malicious URL.
Why this CVE matters
Path traversal flaws look low-impact on paper but routinely chain into full compromise. An attacker who can read arbitrary files often pulls configuration secrets, session databases, or private keys, and many traversal bugs also allow writes that drop a webshell into the document root.
For deployments of pydantic-ai that have been exposed to the public internet during the disclosure window, the operating assumption should be that scanning has already happened. Even where exploitation has not been publicly observed, scanning for the vulnerable fingerprint is cheap and routine. Patching closes the door; log review and credential rotation close out the rest of the response.
Am I affected?
You are affected if your installation matches any of these version ranges:
- pydantic-ai: >= 1.34.0, < 1.51.0
Check your installed version against the list above. If you cannot determine the version, treat the system as affected and follow the upgrade path below.
Open pydantic-ai's About dialog or run the vendor-documented version-check command. Compare the result against the affected ranges in the advisory.
How to fix CVE-2026-25640
- Read the vendor advisory in full: https://github.com/pydantic/pydantic-ai/security/advisories/GHSA-wjp5-868j-wqv7
- Upgrade pydantic-ai to the patched build listed in the vendor advisory.
- Back up the configuration (and database, where applicable) before upgrading.
- Apply the patch in a maintenance window. For HA pairs, upgrade the standby node first, fail over, then upgrade the former primary.
- Restart the affected service so the patched binary loads, then verify the new version (see verification section).
Patched-version commands
Vendor advisory: https://github.com/pydantic/pydantic-ai/security/advisories/GHSA-wjp5-868j-wqv7
Affected: pydantic-ai: >= 1.34.0, < 1.51.0
Patched in: 1.51.0
# Vendor advisory: https://github.com/pydantic/pydantic-ai/security/advisories/GHSA-wjp5-868j-wqv7
# Update inside an existing project.
npm install pydantic-ai@1.51.0
npm audit fix
# Confirm the patched version landed in node_modules.
npm list pydantic-ai
# Lock-file enforcement on CI.
npm ci
# Same workflow from a Windows admin workstation.
npm install pydantic-ai@1.51.0
npm audit fix
npm list pydantic-ai
Verify the fix landed
# Vendor advisory: https://github.com/pydantic/pydantic-ai/security/advisories/GHSA-wjp5-868j-wqv7
# Post-patch verification (replace <service> with the real service unit).
journalctl -u <service> --since "10 minutes ago"
dmesg --since "10 minutes ago"
# Re-scan with your vulnerability scanner (Nessus, Qualys, Tenable, OpenVAS).
# It should no longer flag CVE-2026-25640 on the patched target.
If you cannot patch immediately
Block requests containing ../, ..%2f, or absolute path prefixes at a reverse proxy. Restrict access to the affected endpoint to trusted networks. Apply the patched build as the real fix.
How to verify the fix worked
- After applying the patch, verify the running version in the product's admin UI or via the vendor-documented CLI command.
- Confirm the patched build matches the version listed in the vendor advisory.
- Run an authenticated vulnerability scan with a current signature set and confirm the scanner no longer flags CVE-2026-25640.
- Review logs for the entire pre-patch window for indicators of compromise listed in the vendor or CISA advisory.
- Confirm any network-layer mitigations that were applied as a stopgap have been reverted (or left in place intentionally) once the patch is verified.
If your installation was internet-reachable during the disclosure window, treat log review as part of the remediation rather than an optional follow-up. Look for unusually long URI paths containing traversal sequences, unexpectedly large responses from the affected endpoint, and outbound requests from the application to internal addresses or cloud-metadata endpoints. Treat any sensitive file the bug could disclose as exposed.
Frequently asked questions
Is CVE-2026-25640 being exploited in the wild?
Public exploitation has not been confirmed by CISA at the time of writing. Treat the patch as time-sensitive anyway; reports often lag actual abuse.
Will a WAF or IDS rule fully mitigate CVE-2026-25640?
No. Network-layer filters can reduce noise and slow opportunistic scanners, but they will not stop a determined attacker. The vendor patch is the only durable fix.
How long should I plan for the upgrade?
Typical vendor-documented upgrade windows for pydantic-ai run from a few minutes to under an hour depending on cluster size. Test in a staging environment first and follow the vendor's documented HA upgrade order.
References
- Official vendor advisory: https://github.com/pydantic/pydantic-ai/security/advisories/GHSA-wjp5-868j-wqv7
- NVD entry: https://nvd.nist.gov/vuln/detail/CVE-2026-25640
- CISA KEV catalog: https://www.cisa.gov/known-exploited-vulnerabilities-catalog
- Additional vendor or research reference: https://github.com/pydantic/pydantic-ai/releases/tag/v1.51.0
*This guide was assembled from the official vendor advisory, the NVD record, and the CISA KEV catalog entry on 2026-05-25. Always confirm against the vendor advisory before applying changes in production.*