16,517 results · Clear filters · Feed
-
Medium 6.5
Connection redirection via percent-encoded delimiter injection in connection string hosts
PyMongo's connection string parsing decodes percent-encoded characters in the host portion before the host list is separated on its delimiters. When an application places a hostname value supplied by an unauthenticated party into a connection string, that par…
-
Medium 5.0
Forced local Unix socket connection via dot-sock KMS endpoint in client-side field encryption
The client-side field level encryption support in the MongoDB Python Driver can treat a key management endpoint value ending in ".sock" as a local Unix domain socket path rather than a remote host. A user with write access to the encryption key metadata store…
-
Medium 5.4
Filament: Multi-factor authentication (app) management actions do not require password reauthentication
Filament is a collection of full-stack components for accelerated Laravel development. From 4.0.0 until 4.13.3 and 5.8.3, app-based multi-factor authentication management actions do not consistently require confirmation of the current password. An attacker wi…
-
Medium 6.5
PyJWT accepts inconsistent OKP x/d JWKs, causing public/private key identity confusion
PyJWT is a Python implementation of JSON Web Token standards. From 2.1.0 until 2.15.0, PyJWT OKPAlgorithm.from_jwk in jwt/algorithms.py is affected because private-JWK import path does not compare the public key derived from d with x. This occurs when an OKP …
-
Medium 6.1
DOMPurify before 3.4.16 XSS via IN_PLACE Returned Rawtext Root
DOMPurify before 3.4.16 contains a cross-site scripting vulnerability that allows attackers to bypass sanitization because IN_PLACE mode returns a force-removed style element whose text content still carries markup. Attackers can supply a style element contai…
-
Medium 5.3
vLLM: Qwen2-VL / Qwen3-VL video samplers bound on request-controlled max_frames, which the num_frames ceiling does not reach
vLLM is an inference and serving engine for large language models. From 0.24.0 until 0.30.0, the Qwen2VLVideoBackend and Qwen3VLVideoBackend classes accept request-level values for the media_io_kwargs.video.max_frames and media_io_kwargs.video.fps fields with…
-
Medium 5.3
vLLM: GLMGA video sampling permits request-driven CPU and memory exhaustion
vLLM is an inference and serving engine for large language models. Prior to 0.30.0, a caller can use the request-level media_io_kwargs field to select the GLMGA video backend and supply large values for the fps and max_frames options without a strict work cei…
-
Medium 6.5
vLLM: Scale-out disaggregated multimodal transport trusts caller-supplied features
vLLM is an inference and serving engine for large language models. Prior to 0.30.0, the /inference/v1/generate endpoint in the disaggregated scale-out path accepts caller-supplied tensors in the features.kwargs_data field, cache identifiers in the features.mm…
-
Medium 6.5
vLLM: Structured-output request errors escape the request boundary and terminate the shared EngineCore — engine-fatal denial of service (3 sites)
vLLM is an inference and serving engine for large language models. Prior to 0.30.0, structured-output request failures can escape request-scoped validation and reach the EngineCore fatal-error path. A per-request backend mismatch can re-raise a grammar compil…
-
Medium 4.2
vLLM: Flash late-interaction scoring caches query embeddings under a caller-controlled request id — cross-request integrity break and induced errors on `/score` and `/rerank`
vLLM is an inference and serving engine for large language models. Prior to 0.30.0, flash late-interaction scoring at the /score and /rerank endpoints derives each worker's query_key value from the caller-controlled X-Request-Id header. A concurrent request t…
-
Medium 6.5
vLLM: Loose `cache_salt` validation lets a single request kill EngineCore on LMCache-MP deployments — uncaught downstream `ValueError` denial of service
vLLM is an inference and serving engine for large language models. Prior to 0.30.0, OpenAI-compatible request models accept a non-empty cache_salt value without enforcing the character and length restrictions required by the IPCCacheServerKey consumer in LMCa…
-
Medium 5.9
uv: Path traversal on Windows through wheel extraction
uv is a Python package and project manager written in Rust. From 0.12.7 until 0.12.18, uv wheel extraction on Windows can process a malicious wheel in a way that writes a file outside the installation prefix, including an executable in a directory already pre…
-
Medium 6.5
Mako: Path traversal via drive-letter URI on Windows in TemplateLookup
Mako is a template library written in Python. Prior to 1.4.2, on Windows, TemplateLookup.get_template() in mako/lookup.py resolves template URIs with posixpath, while Template.__init__() in mako/template.py validates them with os.path, which uses ntpath. A UR…
-
Medium 5.3
smol-toml: Quadratic-time parse() from parseKey rescanning to end of document on each key line
### Summary `parse()` has a quadratic-time path in `parseKey`, reachable on default options with ordinary valid input. For every key line and table-header line, `parseKey` (dist/struct.js, lines 58 and 86) finds the dotted-key separator with `ctx.s.indexOf('.…
-
Medium 5.3
Multidict: Reference leak in CIMultiDict/MultiDict items-view union and subtraction
Multidict is an implementation of a multidict data structure. From 6.7.0 until 6.9.1, the C extension's items-view reflected union operation, operand | d.items(), in multidict_itemsview_or2_impl and subtraction operation, d.items() - operand, in multidict_ite…
-
Medium 5.3
Rustls 0.23.13 before 0.23.45 TLS 1.3 Handshake Encryption Level Bypass
Rustls from 0.23.13 before 0.23.45 accepts TLS 1.3 handshake messages at the wrong encryption level when they follow a key-changing message in the same record, because Deframer::aligned does not enforce RFC 8446 record-boundary alignment. A peer or on-path at…
-
Medium 5.3
Rustls: TLS 1.3 handshake messages incorrectly accepted across encryption level boundaries
Rustls accepted TLS 1.3 handshake messages sent at the wrong encryption level when they followed a key-changing message in the same record. The handshake transcript is still authenticated, so a network-position attacker cannot use this to alter or complete a …
-
Medium 5.8
Nx: Path traversal in nx migrate package-migrations extraction
Nx is a monorepo solution for TypeScript and polyglot codebases. From 13.10.0 until 22.7.10 and 23.2.1, Nx migration planning reads the nx-migrations.migrations value from a target package manifest without validating that it is a contained relative path. A ho…
-
Medium 6.5
OpenClaw Channel Plugins before 2026.8.1 Channel Read Allowlist Bypass
OpenClaw channel plugins (@openclaw/msteams, @openclaw/feishu, @openclaw/matrix, and @openclaw/googlechat) before 2026.8.1 do not enforce the configured channel read allowlist for caller-supplied explicit read targets in message, reaction, pin, member, and re…
-
Medium 6.8
vm2 before 3.12.2 Host Process Termination via Construct Trap
vm2 before 3.12.2 does not apply host-side Promise rejection handling in the sandbox-to-host construct trap. In BaseHandler, the apply trap calls markHostPromiseHandled() on the returned value, but the adjacent construct path returns the result of Reflect.con…
-
Medium 5.9
vLLM: Unbounded Prometheus label cardinality from attacker-controlled HTTP method tokens in the vLLM Rust frontend metrics middleware (unauthenticated denial of service)
vLLM is an inference and serving engine for large language models. Prior to 0.30.0, the Rust frontend's track_http_metrics middleware records the raw HTTP method token as a Prometheus label for requests reaching registered routes. An unauthenticated attacker …
-
Medium 6.5
Sensitive information written to logs by Snowflake drivers
Insertion of sensitive information into log files in the Snowflake Python, Go, JDBC, Node.js, PHP PDO, and ODBC drivers allowed authentication tokens, query-result encryption keys, pre-signed cloud-storage URLs, and SAML assertions to be written to diagnostic…
-
Medium 5.3
Hickory DNS 0.25.0 before 0.26.2 Query Amplification via Irrelevant CNAME Records
The hickory-resolver component of Hickory DNS 0.25.0 before 0.26.2 follows irrelevant CNAME records when resolving recursive queries, sending queries that are unnecessary to answer the original request. Attackers controlling DNS responses can include CNAME re…
-
Medium —
hickory-resolver follows irrelevant CNAME records
When the Hickory DNS resolver follows CNAME records, it sends queries that are not necessary to answer the original recursive query. If there are any CNAME records in the authority section or additional section of the response, queries will be sent for those …
-
Medium 5.9
PostCSS: Quadratic complexity in flat selector parsing allows CPU exhaustion
PostCSS Selector Parser is a CSS selector parser that integrates with PostCSS but does not require it. Prior to 7.1.6, src/parser.js splitWord() can receive a flat selector as one word token carrying many class or ID indexes because period and hash characters…
-
Medium 6.3
Angular SSR: Path Traversal to Sibling Directories in CommonEngine on Windows
The Angular SSR is a server-rise rendering tool for Angular applications. Prior to versions 20.3.36, 21.2.23, and 22.1.7, the CommonEngine retrieveSSGPage prerendered-page retrieval logic in @angular/ssr/node, and in @angular/ssr for versions 17 through 18, a…
-
Medium 5.8
Multiple @opentelemetry/instrumentation-* packages expose database username via unconditional db.user span attribute
OpenTelemetry JavaScript Contrib provides instrumentation libraries for collecting telemetry from JavaScript applications. Prior to versions 0.66.0 of @opentelemetry/instrumentation-cassandra-driver, 0.65.0 of @opentelemetry/instrumentation-knex, 0.67.0 of @o…
-
Medium —
fast-copy: Stack exhaustion in fast-copy when copying deeply-nested values
## Summary `fast-copy` traverses values recursively with no bound on depth. Copying a sufficiently deeply-nested value exhausts the JavaScript call stack and throws a native `RangeError: Maximum call stack size exceeded` from inside the library. ## Details Bo…
-
Medium 6.2
stream-json: JSONC parser and verifier re-scan the whole accumulated comment on every input chunk
stream-json is a micro-library of stream components for processing JSON and JSONC with a minimal memory footprint. Prior to 3.6.0, the JSONC parser at stream-json/jsonc/parser.js and verifier at stream-json/jsonc/verifier.js restart comment-terminator scannin…
-
Medium 5.1
stream-json: Prototype pollution: Assembler writes this.current[this.key] on plain objects
stream-json is a micro-library of stream components for processing JSON and JSONC with a minimal memory footprint. Prior to 3.6.0, Assembler materializes object properties with plain assignment, so an input key named __proto__ invokes the inherited setter and…
-
Medium 5.8
vm2 before 3.11.8 Information Disclosure via util.getCallSites
vm2 is a sandbox for running untrusted Node.js code. In versions <= 3.11.7, NodeVM exposes the host `util` module to the sandbox as an unfiltered shallow copy (`Object.assign({}, util)` in `defaultBuiltinLoaderUtil`), and the deprecated `sys` builtin (an alia…
-
Medium 5.8
vm2 3.11.0 before 3.11.7 Information Disclosure via Error Stack
vm2 versions 3.11.0 through 3.11.6 leak absolute host filesystem paths to sandboxed code through error stack formatting. Attacker-supplied code can force the host-realm source transformer to throw a SyntaxError (for example by calling eval with malformed sour…
-
Medium 6.9
Docling: Arbitrary local file read via draw:image xlink:href in the OpenDocument backend
Docling simplifies document processing by parsing diverse formats and providing integrations with the generative AI ecosystem. From 2.107.0 until 2.120.3, docling/backend/opendocument_backend.py uses the xlink:href attribute value of a draw:image element as a…
-
Medium 5.8
@fastify/busboy vulnerable to CRLF injection via multipart Content-Disposition filename and name
@fastify/busboy is a multipart form-data parser for Node.js. Its multipart part-header parser splits header lines only on the two-byte carriage-return line-feed sequence, so a lone carriage return or line feed embedded in a part header is not treated as a lin…
-
Medium 6.3
langflow-ai langflow LambdaFilterComponent lambda_filter.py eval code injection
A weakness has been identified in langflow-ai langflow up to 1.10.2. This affects the function eval of the file src/lfx/src/lfx/components/llm_operations/lambda_filter.py of the component LambdaFilterComponent. Executing a manipulation can lead to code inject…
-
Medium 4.9
SiYuan before v3.8.1 Missing Authorization via /history and /repo/diff
SiYuan before v3.8.1 does not apply the IsForbiddenAbsPath guard (introduced in GHSA-c8r8-95hg-mp34) to the /history/*path and /repo/diff/*path endpoints in kernel/server/serve.go. These routes require admin authentication but construct file paths independent…
-
Medium 6.5
@orpc/zod: Prototype injection in smart coercion
oRPC is a tool that helps build APIs that are end-to-end type-safe and adhere to OpenAPI standards. Prior to 1.14.10, the @orpc/zod ZodSmartCoercionPlugin and experimental_ZodSmartCoercionPlugin collect object and record properties in plain objects and resolv…
-
Medium 5.7
SiYuan before v3.8.1 Path Traversal via asset.upload
SiYuan before v3.8.1 contains a path traversal vulnerability in the asset.upload MCP tool that accepts arbitrary absolute file paths without workspace boundary validation. Attackers can induce the AI Agent to upload sensitive files such as SSH keys or credent…
-
Medium —
sqlite3-ruby: Use-After-Free in SQLite Aggregate Arguments in Heap-Allocated Argument Array
## Summary Using `Database#create_aggregate`, `#create_aggregate_handler`, or `Database#define_aggregator` to define an aggregate function that takes two or more arguments, and then evaluating it over TEXT or BLOB column values, can free the Ruby objects hold…
-
Medium 6.1
Arbitrary JavaScript Execution via openUrl in @a2ui/web_core
The openUrl function in @a2ui/web_core passes an agent-controlled URL directly to window.open() without validating the URI scheme. A malicious agent can supply a javascript: URI as the url argument of a Button component's functionCall action. When the user cl…