CVE-2026-107295
`pydantic-ai-slim` web UI `/api/chat` accepts browser-simple cross-origin requests that can trigger agent tool execution
Pydantic AI is a Python agent framework for building applications and workflows with Generative AI. From 1.34.0 until 1.107.4 and 2.28.0, the Agent.to_web() and clai web development chat endpoint has missing request content-type validation. A website visited by a developer can submit a browser-compatible request to a loopback-hosted chat server, causing the served agent to run and execute tools with the privileges and credentials of the local process; client-relayed approval decisions also leave requires_approval=True tools exposed. Binding to localhost does not prevent a browser page from reaching the loopback address. This issue is fixed in versions 1.107.4 and 2.28.0.
Attack profile
How easy it is to exploit
- Attack vector
- Network
- Complexity
- Low
- Privileges needed
- None
- User interaction
- Required
What an attacker can affect
- Confidentiality
- Low
- Integrity
- High
- Availability
- Low
Exploit likelihood
What each source reports
CVE List Score used
`pydantic-ai-slim` web UI `/api/chat` accepts browser-simple cross-origin requests that can trigger agent tool execution
Pydantic AI is a Python agent framework for building applications and workflows with Generative AI. From 1.34.0 until 1.107.4 and 2.28.0, the Agent.to_web() and clai web development chat endpoint has missing request content-type validation. A website visited by a developer can submit a browser-compatible request to a loopback-hosted chat server, causing the served agent to run and execute tools with the privileges and credentials of the local process; client-relayed approval decisions also leave requires_approval=True tools exposed. Binding to localhost does not prevent a browser page from reaching the loopback address. This issue is fixed in versions 1.107.4 and 2.28.0.
- Record
CVE-2026-107295- CVSS 3.1
- 7.6 High
CVSS:3.1/AV:N/AC:L/PR:N/UI:R/S:U/C:L/I:H/A:L - Weakness
- CWE-352
- Updated
- 2026-11-02
EUVDEUVD-2026-95106
Pydantic AI is a Python agent framework for building applications and workflows with Generative AI. From 1.34.0 until 1.107.4 and 2.28.0, the Agent.to_web() and clai web development chat endpoint has missing request content-type validation. A website visited by a developer can submit a browser-compatible request to a loopback-hosted chat server, causing the served agent to run and execute tools with the privileges and credentials of the local process; client-relayed approval decisions also leave requires_approval=True tools exposed. Binding to localhost does not prevent a browser page from reaching the loopback address. This issue is fixed in versions 1.107.4 and 2.28.0.
- Record
EUVD-2026-95106- CVSS 3.1
- 7.6 High
CVSS:3.1/AV:N/AC:L/PR:N/UI:R/S:U/C:L/I:H/A:L - Updated
- 2026-11-02
GitHub AdvisoriesGHSA-h4xc-3qfq-jf93
Pydantic AI Web chat UI (`Agent.to_web()`, `clai web`): a website visited by the developer can trigger agent runs and server-side tool execution on the local chat endpoint
Summary
The Pydantic AI development web chat UI (Agent.to_web(), clai web) did not check the content type of requests to its chat endpoint. This allowed a website visited by a developer to submit a request to a chat UI running on that developer's machine, causing the served agent to run and to execute its tools with the privileges and credentials of the local process.
Binding the web UI to localhost — the default — does not prevent this, because a page open in the developer's browser can reach the loopback address.
Impact
Applications and developers serving an agent through Agent.to_web() or clai web. The consequences depend on the tools the served agent exposes, and can include data disclosure as well as unwanted tool side effects. Tools marked requires_approval=True were not protected either, because the endpoint trusts approval decisions relayed by the client.
Mitigation
Upgrade to a patched version. The chat endpoint now requires Content-Type: application/json and rejects other requests before the request body is parsed and before the agent runs. The bundled chat UI already sends this header; scripts and other non-browser clients that call the endpoint directly may need to be updated to send it.
If you cannot upgrade, don't run the web UI while browsing untrusted sites, stop it when you aren't using it, and don't serve an agent with side-effecting tools through it.
Credits
- Thai Son Dinh from VinSOC Labs (R&D)
- Record
GHSA-h4xc-3qfq-jf93- CVSS 3.1
- 7.6 High
CVSS:3.1/AV:N/AC:L/PR:N/UI:R/S:U/C:L/I:H/A:L - Weakness
- CWE-346, CWE-352
- Updated
- 2026-11-02
NVD
Pydantic AI is a Python agent framework for building applications and workflows with Generative AI. From 1.34.0 until 1.107.4 and 2.28.0, the Agent.to_web() and clai web development chat endpoint has missing request content-type validation. A website visited by a developer can submit a browser-compatible request to a loopback-hosted chat server, causing the served agent to run and execute tools with the privileges and credentials of the local process; client-relayed approval decisions also leave requires_approval=True tools exposed. Binding to localhost does not prevent a browser page from reaching the loopback address. This issue is fixed in versions 1.107.4 and 2.28.0.
- Record
CVE-2026-107295- CVSS 3.1
- 7.6 High
CVSS:3.1/AV:N/AC:L/PR:N/UI:R/S:U/C:L/I:H/A:L - Weakness
- CWE-352
- Updated
- 2026-11-02
OSV
`pydantic-ai-slim` web UI `/api/chat` accepts browser-simple cross-origin requests that can trigger agent tool execution
Pydantic AI is a Python agent framework for building applications and workflows with Generative AI. From 1.34.0 until 1.107.4 and 2.28.0, the Agent.to_web() and clai web development chat endpoint has missing request content-type validation. A website visited by a developer can submit a browser-compatible request to a loopback-hosted chat server, causing the served agent to run and execute tools with the privileges and credentials of the local process; client-relayed approval decisions also leave requires_approval=True tools exposed. Binding to localhost does not prevent a browser page from reaching the loopback address. This issue is fixed in versions 1.107.4 and 2.28.0.
- Record
CVE-2026-107295- CVSS 3.1
- 7.6 High
CVSS:3.1/AV:N/AC:L/PR:N/UI:R/S:U/C:L/I:H/A:L - Weakness
- CWE-352
- Updated
- 2026-11-02
OSVGHSA-h4xc-3qfq-jf93
Pydantic AI Web chat UI (`Agent.to_web()`, `clai web`): a website visited by the developer can trigger agent runs and server-side tool execution on the local chat endpoint
Summary
The Pydantic AI development web chat UI (Agent.to_web(), clai web) did not check the content type of requests to its chat endpoint. This allowed a website visited by a developer to submit a request to a chat UI running on that developer's machine, causing the served agent to run and to execute its tools with the privileges and credentials of the local process.
Binding the web UI to localhost — the default — does not prevent this, because a page open in the developer's browser can reach the loopback address.
Impact
Applications and developers serving an agent through Agent.to_web() or clai web. The consequences depend on the tools the served agent exposes, and can include data disclosure as well as unwanted tool side effects. Tools marked requires_approval=True were not protected either, because the endpoint trusts approval decisions relayed by the client.
Mitigation
Upgrade to a patched version. The chat endpoint now requires Content-Type: application/json and rejects other requests before the request body is parsed and before the agent runs. The bundled chat UI already sends this header; scripts and other non-browser clients that call the endpoint directly may need to be updated to send it.
If you cannot upgrade, don't run the web UI while browsing untrusted sites, stop it when you aren't using it, and don't serve an agent with side-effecting tools through it.
Credits
- Thai Son Dinh from VinSOC Labs (R&D)
- Record
GHSA-h4xc-3qfq-jf93- CVSS 3.1
- 7.6 High
CVSS:3.1/AV:N/AC:L/PR:N/UI:R/S:U/C:L/I:H/A:L - Weakness
- CWE-346, CWE-352
- Updated
- 2026-11-02
OSVPYSEC-2026-4278
Pydantic AI Web chat UI (`Agent.to_web()`, `clai web`): a website visited by the developer can trigger agent runs and server-side tool execution on the local chat endpoint
Summary
The Pydantic AI development web chat UI (Agent.to_web(), clai web) did not check the content type of requests to its chat endpoint. This allowed a website visited by a developer to submit a request to a chat UI running on that developer's machine, causing the served agent to run and to execute its tools with the privileges and credentials of the local process.
Binding the web UI to localhost — the default — does not prevent this, because a page open in the developer's browser can reach the loopback address.
Impact
Applications and developers serving an agent through Agent.to_web() or clai web. The consequences depend on the tools the served agent exposes, and can include data disclosure as well as unwanted tool side effects. Tools marked requires_approval=True were not protected either, because the endpoint trusts approval decisions relayed by the client.
Mitigation
Upgrade to a patched version. The chat endpoint now requires Content-Type: application/json and rejects other requests before the request body is parsed and before the agent runs. The bundled chat UI already sends this header; scripts and other non-browser clients that call the endpoint directly may need to be updated to send it.
If you cannot upgrade, don't run the web UI while browsing untrusted sites, stop it when you aren't using it, and don't serve an agent with side-effecting tools through it.
Credits
- Thai Son Dinh from VinSOC Labs (R&D)
- Record
PYSEC-2026-4278- CVSS 3.1
- 7.6 High
CVSS:3.1/AV:N/AC:L/PR:N/UI:R/S:U/C:L/I:H/A:L - Updated
- 2026-11-02
OSVPYSEC-2026-4287
Pydantic AI Web chat UI (`Agent.to_web()`, `clai web`): a website visited by the developer can trigger agent runs and server-side tool execution on the local chat endpoint
Summary
The Pydantic AI development web chat UI (Agent.to_web(), clai web) did not check the content type of requests to its chat endpoint. This allowed a website visited by a developer to submit a request to a chat UI running on that developer's machine, causing the served agent to run and to execute its tools with the privileges and credentials of the local process.
Binding the web UI to localhost — the default — does not prevent this, because a page open in the developer's browser can reach the loopback address.
Impact
Applications and developers serving an agent through Agent.to_web() or clai web. The consequences depend on the tools the served agent exposes, and can include data disclosure as well as unwanted tool side effects. Tools marked requires_approval=True were not protected either, because the endpoint trusts approval decisions relayed by the client.
Mitigation
Upgrade to a patched version. The chat endpoint now requires Content-Type: application/json and rejects other requests before the request body is parsed and before the agent runs. The bundled chat UI already sends this header; scripts and other non-browser clients that call the endpoint directly may need to be updated to send it.
If you cannot upgrade, don't run the web UI while browsing untrusted sites, stop it when you aren't using it, and don't serve an agent with side-effecting tools through it.
Credits
- Thai Son Dinh from VinSOC Labs (R&D)
- Record
PYSEC-2026-4287- CVSS 3.1
- 7.6 High
CVSS:3.1/AV:N/AC:L/PR:N/UI:R/S:U/C:L/I:H/A:L - Updated
- 2026-11-02
Affected products
Affected packages
| Package | Affected | Fixed |
|---|---|---|
pydantic-aiPyPI |
1.34.0 up to 1.107.4 | 1.107.4 |
pydantic-aiPyPI |
2.0.0b1 up to 2.28.0 | 2.28.0 |
pydantic-ai-slimPyPI |
1.34.0 up to 1.107.4 | 1.107.4 |
pydantic-ai-slimPyPI |
2.0.0b1 up to 2.28.0 | 2.28.0 |
pydantic-aiPyPI |
1.34.0 up to 1.107.4, plus listed versions 1.100.01.101.01.102.01.103.01.104.0 Show 45 moreShow less1.105.01.106.01.107.01.107.11.107.21.35.01.36.01.37.01.38.01.39.01.39.11.40.01.41.01.42.01.43.01.44.01.46.01.47.01.48.01.49.01.50.01.51.01.52.01.53.01.54.01.55.01.56.01.57.01.58.01.59.01.60.01.61.01.62.01.63.01.64.01.65.01.66.01.67.01.68.01.69.01.70.01.71.01.72.01.73.01.74.0 |
1.107.4 |
pydantic-aiPyPI |
2.0.0b1 up to 2.28.0, plus listed versions 2.0.02.0.0b22.0.0b32.0.0b42.0.0b5 Show 33 moreShow less2.0.0b62.0.0b72.1.02.10.02.11.02.12.02.13.02.14.02.14.12.15.02.16.02.17.02.18.02.19.02.2.02.20.02.21.02.22.02.23.02.24.02.25.02.26.02.27.02.27.12.3.02.4.02.5.02.5.12.6.02.7.02.8.02.9.02.9.1 |
2.28.0 |
pydantic-ai-slimPyPI |
1.34.0 up to 1.107.4, plus listed versions 1.100.01.101.01.102.01.103.01.104.0 Show 45 moreShow less1.105.01.106.01.107.01.107.11.107.21.35.01.36.01.37.01.38.01.39.01.39.11.40.01.41.01.42.01.43.01.44.01.46.01.47.01.48.01.49.01.50.01.51.01.52.01.53.01.54.01.55.01.56.01.57.01.58.01.59.01.60.01.61.01.62.01.63.01.64.01.65.01.66.01.67.01.68.01.69.01.70.01.71.01.72.01.73.01.74.0 |
1.107.4 |
pydantic-ai-slimPyPI |
2.0.0b1 up to 2.28.0, plus listed versions 2.0.02.0.0b22.0.0b32.0.0b42.0.0b5 Show 33 moreShow less2.0.0b62.0.0b72.1.02.10.02.11.02.12.02.13.02.14.02.14.12.15.02.16.02.17.02.18.02.19.02.2.02.20.02.21.02.22.02.23.02.24.02.25.02.26.02.27.02.27.12.3.02.4.02.5.02.5.12.6.02.7.02.8.02.9.02.9.1 |
2.28.0 |
pydantic-aiPyPI |
1.34.0 up to 1.107.4, 2.0.0b1 up to 2.28.0, plus listed versions 1.100.01.101.01.102.01.103.01.104.0 Show 45 moreShow less1.105.01.106.01.107.01.107.11.107.21.35.01.36.01.37.01.38.01.39.01.39.11.40.01.41.01.42.01.43.01.44.01.46.01.47.01.48.01.49.01.50.01.51.01.52.01.53.01.54.01.55.01.56.01.57.01.58.01.59.01.60.01.61.01.62.01.63.01.64.01.65.01.66.01.67.01.68.01.69.01.70.01.71.01.72.01.73.01.74.0 |
1.107.42.28.0 |
pydantic-ai-slimPyPI |
1.34.0 up to 1.107.4, 2.0.0b1 up to 2.28.0, plus listed versions 1.100.01.101.01.102.01.103.01.104.0 Show 45 moreShow less1.105.01.106.01.107.01.107.11.107.21.35.01.36.01.37.01.38.01.39.01.39.11.40.01.41.01.42.01.43.01.44.01.46.01.47.01.48.01.49.01.50.01.51.01.52.01.53.01.54.01.55.01.56.01.57.01.58.01.59.01.60.01.61.01.62.01.63.01.64.01.65.01.66.01.67.01.68.01.69.01.70.01.71.01.72.01.73.01.74.0 |
1.107.42.28.0 |
History
| Recorded | Changed | Sources |
|---|---|---|
| first indexed | CVE List, EUVD, GitHub Advisories, NVD, OSV |
Dates are when Every CVE recorded each change. Upstream dates may differ.
In short
CVE-2026-107295 is a high vulnerability (CVSS 7.6) in Pydantic Pydantic-ai, published October 8, 2026. No source lists public exploit code. EPSS puts the chance of exploitation in the next 30 days at 0.16%, the 4th percentile. Fixed in pydantic-ai 1.107.4.
References
github.com/pydantic/pydantic-ai/security/advisories/GHSA-h4xc-3qfq-jf93
Show 7 moreShow less
nvd.nist.gov/vuln/detail/CVE-2026-107295
pypi.org/project/pydantic-ai
Left out here: 200 versions. The JSON record has them.