Security Vulnerabilities
Pydantic AI is a Python agent framework for building applications and workflows with Generative AI. From 1.34.0 until 2.30.0, the Agent.to_web() and clai web development chat server does not validate the Host header, allowing a website visited by a developer to use DNS rebinding to reach a loopback-hosted agent as a same-origin service. The hostile page can read the served UI and submit chat requests that execute agent tools with the local process's privileges and credentials, causing data disclosure or unwanted side effects. Binding to localhost, Origin checks, and CSRF tokens do not prevent the same-origin DNS rebinding path. This issue is fixed in versions 1.107.5 and 2.30.0.
CVSS Score
6.4
EPSS Score
0.001
Published
2026-10-08
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.
CVSS Score
2.3
EPSS Score
0.003
Published
2026-10-08
Pydantic AI is a Python agent framework for building applications and workflows with Generative AI. From 1.77.0 until 1.107.6 and 2.44.0, the local web_fetch_tool and the WebFetch local fallback compare blocked_domains entries with a URL hostname before both values are normalized to the form used by getaddrinfo. An attacker-influenced model can use an equivalent IDNA spelling, non-ASCII label separator, case variation, or trailing root label that resolves to a blocked host but does not match the configured string, causing the application to fetch that host with its own privileges. allowed_domains fails closed for unmatched spellings, and private-IP and cloud-metadata protections remain effective. This issue is fixed in versions 1.107.6 and 2.44.0.
CVSS Score
3.7
EPSS Score
0.004
Published
2026-10-08
Pydantic AI is a Python agent framework for building applications and workflows with Generative AI. From 1.56.0 until 1.107.6 and 2.44.0, applications that opt attacker-influenced URLs into local network access through FileUrl with force_download='allow-local' or web_fetch_tool with allow_local_urls=True can bypass the cloud-metadata blocklist by appending an IPv6 zone identifier to an IPv6 metadata address. IPv6Address equality and hashing include the zone identifier, so the blocklist comparison fails even though the network stack ignores the zone on a non-link-local destination and reaches the metadata service, potentially exposing cloud IAM credentials. The opt-in settings are disabled by default, and the issue requires an IPv6-enabled environment. This issue is fixed in versions 1.107.6 and 2.44.0.
CVSS Score
6.8
EPSS Score
0.003
Published
2026-10-08
Pydantic AI is a Python agent framework for building applications and workflows with Generative AI. From 1.77.0 until 1.107.7 and 2.52.0, the local web_fetch_tool and the WebFetch local fallback can consume excessive CPU and memory during HTML-to-Markdown conversion of attacker-controlled HTML containing deeply nested block elements. Conversion repeatedly reprocesses accumulated text and can greatly expand intermediate output before the returned-content limit is applied, allowing a model-directed fetch to delay other work in the process. Provider-native web fetching is not affected. This issue is fixed in versions 1.107.7 and 2.52.0.
CVSS Score
6.5
EPSS Score
0.003
Published
2026-10-08
Pydantic AI is a Python agent framework for building applications and workflows with Generative AI. From 2.10.0 until 2.53.0, streamed requests made through ConcurrencyLimitedModel or limit_model_concurrency can retain shared concurrency slots because anyio.CapacityLimiter associates an acquired slot with the borrowing task while streaming cleanup can run in a different task. Early stream termination, cancellation, consumer exceptions, or complete stream_text() consumption with debounce_by=0.1 can therefore leave capacity occupied, eventually preventing later requests that share the long-lived limiter from proceeding and causing a denial of service. Agent-level max_concurrency and non-streaming model requests are not affected. This issue is fixed in version 2.53.0.
CVSS Score
7.5
EPSS Score
0.004
Published
2026-10-08
A stack-based buffer overflow vulnerability exists in the diagnostic execution utility of Brocade Fabric OS versions before 10.0.1. When processing command arguments for diagnostic operations, the utility tokenizes user-supplied input into an internal argument array without enforcing boundary checks on the maximum array capacity. An authenticated user with administrative access can exploit this vulnerability by supplying a crafted diagnostic command string containing an excessive number of tokenized arguments. This leads to a memory overwrite resulting in a denial of service (process crash).
CVSS Score
6.9
EPSS Score
0.002
Published
2026-10-08
A missing authorization check in Brocade Fabric OS versions before 10.0.1 REST API interface of affected platform releases allows an authenticated user, regardless of their assigned role or administrative scope, to retrieve complete Monitoring and Alerting Policy Suite (MAPS) violation data across all logical switches. An attacker with low-privilege API access can leverage this endpoint to dump chassis-wide system health metrics, port performance violations, and configuration data without appropriate privileges.
CVSS Score
5.1
EPSS Score
0.001
Published
2026-10-08
An authorization logic vulnerability exists in the Brocade Fabric OS versions before 10.0.1 REST API gateway. The internal gate guarding restricted management endpoints relies exclusively on client-controlled HTTP headers. An authenticated user with any valid REST session can spoof these headers to gain unauthorized access to internal management endpoints. This allows low-privilege users to view sensitive chassis metadata, hardware memory patrolling state, and firmware integrity audit logs.
CVSS Score
5.1
EPSS Score
0.001
Published
2026-10-08
An out-of-bounds memory read vulnerability exists in the web management daemon of Brocade Fabric OS versions before 10.0.1. Unauthenticated HTTP endpoints process specific URL query parameters without validating array index boundaries or performing numerical range checks. An unauthenticated remote attacker can exploit this issue by sending a single, crafted HTTP request containing extreme numerical values in the query string. This causes an invalid memory dereference, resulting in a crash of the web management process (Denial of Service) and potential temporary management-plane disruption.
CVSS Score
7.1
EPSS Score
0.002
Published
2026-10-08