The internet is broken right now.
68 critical CVEs without a fix in the last 48 hours. Somewhere a staging server is public. It always is. 151 more already have a fix that someone still has to apply.
-
CVE-2026-18379
A flaw was found in yggdrasil-worker-package-manager. An authenticated attacker can exploit this issue to achieve arbitrary file modification or potential privilege escalation. The vulnerability exists because repository management operations fail to sanitize directory traversal sequences (such as '../') within repository names. By supplying a crafted repository name, an attacker can manipulate the root-level worker process into creating or deleting files outside the intended directory, compromising host system integrity.
-
CVE-2026-129296
mail-parser is a tool that parses emails by enhancing the Python standard library and extracting message details into a comprehensive object. From 3.8.0 until 4.6.0, the write_attachments() and write_sample() APIs use attacker-controlled attachment filenames without sufficient sanitization and path containment validation. Directory traversal sequences and absolute filesystem paths can cause an attachment to be written outside the intended output directory. An attacker can overwrite files accessible to the parsing process when an application extracts attachments from a crafted message. This issue is fixed in version 4.6.0.
-
CVE-2026-129266
External control of a file path in Amazon Deep Java Library (DJL) versions 0.7.0 through 0.38.0 might allow a remote unauthenticated third party to read arbitrary files that the library process can access. To do this, the third party supplies a crafted model archive. To remediate this issue, users should upgrade to version 0.39.0 or later.
-
CVE-2026-126264
SiYuan before 3.8.9 contains an information disclosure vulnerability that allows authenticated Reader users to retrieve credential-named widget files by bypassing the sensitive-path guard in getFile. Attackers can request paths like /data/widgets/<package>/credentials.json for publish-referenced widgets to obtain API keys, tokens, or other stored secrets.
-
CVE-2026-127031
In Eclipse AAS4J before 3.0.1, `XmlSchemaValidator.validateSchema()` (and `AASXValidator.validateSchema()`, which calls it) validates untrusted XML with a default JAXP `Validator` that resolves external DTDs and external entities. An attacker who supplies a crafted AAS XML document or AASX package can read local files from the host, send their contents out-of-band to an attacker-controlled server, and make the validating process send arbitrary outbound requests (SSRF). Validation still returns no errors. Details: - Vulnerable code: `dataformat-xml/src/main/java/org/eclipse/digitaltwin/aas4j/v3/dataformat/xml/XmlSchemaValidator.java`. `schema.newValidator().validate(new StreamSource(...))` runs without `ACCESS_EXTERNAL_DTD` / `ACCESS_EXTERNAL_SCHEMA` restrictions and without `disallow-doctype-decl`. - Second entry point: `dataformat-aasx/src/main/java/org/eclipse/digitaltwin/aas4j/v3/dataformat/aasx/AASXValidator.java`. The AAS XML part of an `.aasx` package is passed unchanged to the vulnerable validator. - `XmlDeserializer` (Jackson/Woodstox) is not affected, because DTD support is disabled by default there. - Present since the initial import of the code base (commit `cc9da52`). - Fix: harden the validator by setting `ACCESS_EXTERNAL_DTD` and `ACCESS_EXTERNAL_SCHEMA` to `""` and rejecting `DOCTYPE` declarations. The bundled `AAS.xsd` does not need DTDs, so no functionality is lost.
-
CVE-2026-126354
Addressable is an alternative implementation to the URI implementation that is part of Ruby's standard library. From 2.2.5 until 2.9.1, Addressable::URI.normalize_path can resolve a path containing leading parent segments and later parent segments to a different path than an RFC 3986 resolver. URI#normalize, URI#join, URI#==, and URI#=== inherit the result. An application that authorizes the Addressable-normalized path but sends the original URI to another resolver can allow access outside the intended path allowlist. Applications that use the same normalized value for authorization and the request, or reject parent segments, are unaffected. This issue is fixed in version 2.9.1.
-
CVE-2026-126295
NLTK from 3.10.0 before 3.10.4 contains an arbitrary file read vulnerability in BllipParser that allows attackers to open files outside NLTK data roots by supplying unvalidated model paths. Attackers controlling parser_model, reranker_features, reranker_weights, or from_unified_model_dir() input can make the bllipparser C extension read sensitive files or block on FIFOs and devices.
-
CVE-2026-127119
Unverified ownership of the default cluster log location in the EMR module in AWS SDK for pandas (awswrangler) from 1.1.0 before 3.18.0 might allow remote actors to read Amazon EMR cluster logs, including job step output, via a pre-created S3 bucket. To remediate this issue, users should upgrade to version 3.18.0 or later. Users who can't upgrade immediately should pass an explicit log location, or create the default EMR log bucket in their own account in each region they use.
-
CVE-2026-126302
NLTK from 3.9.4 before 3.10.4 contains a symlink following vulnerability in nltk.sem.chat80.sql_query() that bypasses the pathsec data-root sandbox containment check. Attackers who can write into a trusted NLTK data root can plant a symbolic link to read SQLite databases outside the sandbox.
-
CVE-2026-127214
Flowise through 3.1.4 contains a server-side request forgery vulnerability that allows authenticated attackers with chatflows or agentflows create/update permissions to bypass the Custom Function sandbox's SSRF protection by requiring a module subpath inside an allowlisted package, such as axios/dist/node/axios.cjs or node-fetch/lib/index.js, which evades the vm2 mock table and returns an unwrapped HTTP client that skips the checkDenyList validation. Attackers can issue arbitrary GET and POST requests with attacker-controlled bodies via POST /api/v1/node-custom-function to reach loopback and internal-network addresses or the cloud metadata endpoint, enabling network mapping, credential disclosure, and state changes on internal services.
-
CVE-2026-63415
Jane is a set of PHP libraries that generates models and API clients from JSON Schema and OpenAPI specifications. Prior to 7.12.0, Jane's Reference.php::doResolve passes an attacker-controlled OpenAPI $ref target to file_get_contents without restricting its scheme or resolved local path. When a victim generates code from an attacker-influenced document, an HTTP or HTTPS reference can cause requests to attacker-chosen or internal services, while an absolute or parent-relative path can read local JSON or YAML files and inline the retrieved content into generated output. Exploitation occurs during generation or CI processing and can disclose local files and internal-service responses. This issue is fixed in version 7.12.0.
-
CVE-2026-77295
MyMangaDB is a self-hosted manga collection manager with automatic metadata, MyAnimeList import, Docker support, and collection statistics. Prior to 2.7.2, the unauthenticated /api/v1/database/export handler in backend/app/api/v1/endpoints/database.py returns a ZIP containing the live SQLite database and the cover-image library. The database includes the users table and hashed_password values, allowing a remote attacker to obtain all application data and conduct offline password attacks without a session or token. This issue is fixed in version 2.7.2.
-
CVE-2026-115622
A flaw was found in ansible-core. When a task or its parameters are marked no_log: true, sensitive values passed to a module's warn() or deprecate() calls are not redacted and are printed in cleartext to stdout/stderr in warning and deprecation messages. The 'warnings' and 'deprecations' keys are retained in the PRESERVE allow-list in lib/ansible/_internal/_task.py when the task result is censored, so sensitive parameters bypass masking and can be disclosed in output captured in terminal sessions, CI logs, and any automation wrapping ansible-playbook, even though every task touching the secrets is marked no_log: true.
-
CVE-2026-130156
Bazarr is a companion application to Sonarr and Radarr. Prior to 1.6.3-beta.11, the /api/subtitles/subtitles-info handler uses SubtitleNameContents.get() in subtitles_contents.py and reads the caller-supplied subtitlePath after checking only its subtitle extension; because the resolved path is not confined to files indexed by Bazarr, traversal can read files outside the media library. The /api/subtitles/subtitles handler in subtitles.py passes caller-supplied path and reference to sync and translation without per-path authorization, allowing an out-of-library reference file to be read and output to overwrite a file outside the media library. The authenticate decorator gates both endpoints when API authentication is enabled, while auth = None allows access without credentials. Both paths can affect only files that the Bazarr process can read or write. This issue is fixed in version 1.6.3-beta.11.
-
CVE-2023-54417
Nuxt 3.0.0 before 3.4.2 contains a semi-arbitrary file read vulnerability in the vite-node dev server middleware that allows unauthenticated attackers to read files by supplying an absolute path as the moduleId. Attackers can send GET requests to /__nuxt_vite_node__/module/ with an absolute file path to read sensitive system files. Exploitation evidence was first observed by the Shadowserver Foundation on 2023-10-19.
-
CVE-2026-130491
In Eclipse versions 1.5-1.6.13, 2.0.0-2.0.22 and 2.1.0-2.1.2, the client library libmosquitto does not reject Will messages with payloads greater than 65535 bytes when calling `mosquitto_will_set()` or `mosquitto_will_set_v5()`. This results in up to 65535 bytes of uninitialised heap memory being transmitted as the Will. This can be mitigated by only passing payloads with a maximum length of 65535 bytes.