VAITP

Banks generates meaningful LLM prompts using a simple template language. In versions prior to 2.4.4, all four media filters (image, audio, video, document) in banks accept untrusted user input as file paths via Path(value) and pass them directly to open(file_path, "rb") without any path sanitization, canonicalization, or directory restriction. An attacker who controls template variables passed to a banks Prompt can use path traversal (../) to read arbitrary files accessible to the Python process—including .env files, SSH keys, cloud credentials, source code, /etc/passwd, and /etc/shadow—with the content returned base64-encoded in the rendered prompt output, making exfiltration trivial. This is particularly dangerous for applications that use banks to process user-provided template variables before sending prompts to an LLM. This issue has been fixed in version 2.4.4.

PIA's `POST /v1/upload/sbom` endpoint accepts a Bearer JWT and checks its **unverified** `iss` claim against an issuer allowlist using Python's `urlparse` before performing OIDC discovery with `requests`. Because `urlparse` and `requests`/`urllib3` parse an authority string containing a backslash (e.g. `https://attacker-host\@ci.eclipse.org/`) into *different* hostnames, an attacker can craft an issuer that passes the allowlist check yet drives `requests` — and subsequently `urllib.request.urlopen` for JWKS retrieval — to connect to an arbitrary attacker-chosen host, port, and scheme.

Pydantic AI is a Python agent framework for building applications and workflows with Generative AI. In versions 1.88.0 up to but not including 1.107.1 and 2.0.0b1 up to but not including 2.5.0, the UI adapters (AG-UI via Agent.to_ag_ui()/AGUIAdapter, and Vercel AI via VercelAIAdapter) use sanitize_messages to strip unresolved ("dangling") client-submitted tool calls from untrusted message history before it reaches the agent, a defense-in-depth default that prevents the agent from executing tool calls the model never emitted. However, the strip anchored to a message index computed before sanitization ran, so when a trailing client message sanitized to empty and was dropped (for example a client system message under the default manage_system_prompt='server'), a preceding assistant response carrying an unresolved tool call became the new tail and was dispatched without inspection. As a result, a remote client could cause a registered, non-approval server tool to run with client-supplied arguments rather than arguments the model produced. The impact is bounded by what the affected tools do and is most significant for applications that gate tool execution in a model-request hook (before_model_request / after_model_request), since a forged call skips the model turn and bypasses that guardrail; approval-gated tools (requires_approval=True) are not auto-executed by this path. This issue has been fixed in versions 1.107.1 and 2.5.0.

Pydantic AI is a Python agent framework for building Generative AI applications. In versions 1.65.0 through 1.105.0, and 2.0.0b1 through 2.0.0b5, a client that submits message history to a Pydantic AI UI adapter (such as the Vercel AI adapter) can reference arbitrary files in the application's model-provider or cloud-storage account. While file URL parts are validated against a scheme allowlist, UploadedFile references — which point to a file by provider file ID or cloud-storage URI (e.g. s3://…, gs://…) — were forwarded without validation. Because the provider resolves an UploadedFile using the server-side identity (IAM role, service account, or provider API key) rather than the client's, an attacker can craft message history to make the server read objects from its own account or other tenants, given a referenceable identifier. Exploitation requires a valid file identifier, which is not always unguessable depending on how the application names objects. This issue has been fixed in versions 1.106.0 and 2.0.0b6.

Pydantic AI is a Python agent framework for building Generative AI applications. In versions 1.56.0 through 1.98.0, when an application opts a URL into force_download='allow-local' (disabling the default block on private/internal IPs), the cloud-metadata blocklist could be bypassed by encoding the metadata IP in an IPv6 transition form (IPv4-mapped IPv6, 6to4, or NAT64), exposing cloud IAM short-term credentials on dual-stack or translated networks. This is an incomplete fix of GHSA-2jrp-274c-jhv3 / CVE-2026-25580, whose remediation did not hold for IPv6-encoded forms of the metadata IPs. An application is affected only if it explicitly opts a FileUrl (ImageUrl, AudioUrl, VideoUrl, DocumentUrl) into force_download='allow-local' on a URL influenced by untrusted input; it is not affected when using bundled integrations to ingest user input (Agent.to_web / clai web, VercelAIAdapter, AGUIAdapter / Agent.to_ag_ui), since they do not propagate force_download from external data, nor when downloading only from developer-controlled URLs. This issue has been fixed in version 1.99.0.

Linuxfabrik monitoring-plugins provides Python monitoring plugins for Icinga, Nagios, and related monitoring systems. In 6.0.0 and earlier, the redfish-* plugins built request URLs by concatenating an operator-supplied base URL with response-supplied @odata.id links, allowing a malicious or compromised BMC to redirect authenticated Redfish requests and disclose X-Auth-Token or HTTP Basic credentials.

linuxfabrik-lib provides Python modules for database access, caching, shell execution, and API integrations. Prior to version 6.0.0, lib.url.fetch() followed cross-origin redirects while forwarding caller-supplied credential headers other than Authorization and Cookie, allowing a malicious redirect-capable server to receive headers such as X-Auth-Token from authenticated monitoring requests. This issue is fixed in version 6.0.0.

Linuxfabrik monitoring-plugins provides Python monitoring plugins for Icinga, Nagios, and related monitoring systems. In version 6.0.0, the logfile check legacy database migration moved a predictable path from /tmp with os.rename() and allowed a local user controlling the plugin account to place a symlink that would be followed by sqlite3.connect() during a root-run check.

joserfc is a Python library that provides an implementation of several JSON Object Signing and Encryption (JOSE) standards. in versions 1.7.1 and prior, joserfc accepts JWTs with trailing padding (==) which are not conforming to the JOSE specifications. This leads to malleability of the JWTs when consumed by joserfc. Depending on this application this might or not be an issue. This could lead to bypass of token revocation or anti-replay protection when implemented as a deny list of tokens or a deny list of token hashes. Note that ECDSA JWS are always malleable because of the malleability of ECDSA signatures (first test case in the code bellow). This makes a scheme which assumes that JWTs are not malleable brittle. However for other signatures (or MAC) schemes it might make sense to assume non malleability of the token. This issue has been fixed in version 1.7.2.

datamodel-code-generator generates Pydantic v2 models, dataclasses, TypedDict, and msgspec.Struct from OpenAPI, JSON Schema, GraphQL, Avro, Protobuf, and raw JSON, YAML, or CSV. From 0.11.6 until 0.64.0, datamodel-code-generator allows attacker-controlled x-python-import or customTypePath schema extensions to reach src/datamodel_code_generator/parser/jsonschema.py and generated import handling through Import.from_full_path and Imports.create_line in src/datamodel_code_generator/imports.py, allowing a newline to break out of an import statement and execute Python code when the generated model is imported. This issue is fixed in version 0.64.0.