🚨 CVE-2026-82755
Use of Cache Containing Sensitive Information vulnerability in ash-project ash_authentication_oauth2_server allows a shared HTTP cache to serve one tenant's OAuth discovery metadata to another tenant's clients.

The RFC 8414 and RFC 9728 metadata endpoints in AshAuthentication.Phoenix.Oauth2Server.ProtocolRouter return tenant-specific values (issuer, authorization_endpoint, token_endpoint, jwks_uri) when a tenant is set, but sent them with Cache-Control: public, max-age=3600 and no Vary. When the tenant is derived from something other than the URL (a header or the Host) and a shared cache sits in front, the cache key is the URL alone, so a stored response for one tenant is served to another for up to an hour. Affected clients may then send authorization codes and secrets to the wrong tenant's token endpoint and validate tokens against the wrong keys.

This issue affects ash_authentication_oauth2_server: from 0.1.3 before 0.3.1.

🎖@cveNotify

🚨 CVE-2026-82756
Improper Encoding or Escaping of Output vulnerability in ash-project ash_authentication_oauth2_server allows an unauthenticated attacker to inject arbitrary authentication parameters into the WWW-Authenticate challenge header.

BearerPlug and RequireScopePlug built the Bearer resource_metadata="..." challenge by interpolating a resource_metadata URL derived from the request tenant directly into the quoted value. In a multi-tenant application that sets the Ash tenant from request-controlled data (a subdomain, the Host, a path segment, or a header), a tenant containing a " closes the quoted value and appends attacker-chosen auth-params, including a second resource_metadata URL pointing at an attacker-controlled authorization server that spec-following clients follow. Carriage returns and line feeds are rejected by Plug, so this is parameter injection within one header, not response splitting.

This issue affects ash_authentication_oauth2_server: from 0.1.3 before 0.3.1.

🎖@cveNotify

🚨 CVE-2026-82757
Server-Side Request Forgery (SSRF) vulnerability in ash-project ash_authentication_oauth2_server allows an attacker who controls a client metadata URL and its DNS to make the server connect to internal or loopback addresses.

public_ip?/1 in AshAuthentication.Oauth2Server.CIMD.ReqFetcher enforces the outbound policy for CIMD metadata fetches. It classified several address forms as publicly routable that are not: IPv4-compatible ::/96 (for example ::127.0.0.1), SIIT IPv4-translated ::ffff:0:0:0/96, and deprecated site-local fec0::/10. A returned AAAA record in one of these ranges passed the policy, so a fetch pinned to that address reached space the policy was meant to block.

This issue affects ash_authentication_oauth2_server: from 0.3.0 before 0.3.1.

🎖@cveNotify

🚨 CVE-2026-82758
Improper Authentication vulnerability in ash-project ash_authentication_oauth2_server allows an unauthenticated attacker to register OAuth clients even when Dynamic Client Registration is gated by an initial access token.

resolve_secret/3 in AshAuthentication.Oauth2Server (reached through __resolve_secret__!) treated any return other than {:ok, _} or :error from a configured {module, function, args} or 2-arity-function secret provider as a valid secret, wrapping nil, false, or "" as {:ok, value}. When the initial_access_token resolves to such an empty value, POST /oauth/register compares the presented bearer token against it and the comparison passes with no token supplied, so registration is open although it was configured closed. The same fail-open affected other resolved secrets such as signing_secret.

This issue affects ash_authentication_oauth2_server: from 0.1.0 before 0.3.1.

🎖@cveNotify

🚨 CVE-2026-86439
knowns versions before 0.30.0 fail to validate filesystem paths in MCP tool arguments, allowing attackers to read, create, overwrite and delete files outside the project directory. Attackers can supply path arguments containing directory traversal sequences to access arbitrary Markdown files accessible to the server process.

🎖@cveNotify

🚨 CVE-2026-86538
knowns versions before 0.30.0 contain a path traversal vulnerability in the POST /api/templates/preview endpoint that allows unauthenticated attackers to read arbitrary files. Attackers can supply directory traversal sequences in the templateFile parameter to bypass path restrictions and read sensitive files like credentials and configuration through the JSON response.

🎖@cveNotify

🚨 CVE-2026-86539
knowns through 0.33.0 contains a server-side request forgery vulnerability in the POST /api/embedding-models/test endpoint that issues outbound requests to caller-supplied destinations without validation. Attackers can enumerate internal hosts and cloud metadata endpoints by observing transport error messages that reveal network reachability information.

🎖@cveNotify

🚨 CVE-2026-86540
knowns versions before 0.30.0 fail to validate the settings.lsp.languages binary field in project configuration files, allowing attackers to execute arbitrary binaries by crafting a malicious .knowns/config.json file. When a repository with a crafted configuration is opened, the unvalidated binary path is executed twice under the user's account without any verification.

🎖@cveNotify

🚨 CVE-2026-86541
knowns versions before 0.30.0 contain a path traversal vulnerability in the handleCodeReplace() function that allows attackers to overwrite arbitrary files outside the project root. Attackers can supply absolute paths or relative paths containing directory traversal sequences to write malicious content to sensitive files like shell startup scripts or SSH configuration files.

🎖@cveNotify

🚨 CVE-2026-86542
knowns before 0.30.0 fails to validate import names in the import routes, allowing unauthenticated attackers to write files outside the imports directory. Attackers can supply traversal sequences in the name parameter to escape the imports directory and overwrite arbitrary files writable by the server process.

🎖@cveNotify

🚨 CVE-2026-86543
knowns versions before 0.30.0 serve the management API without authentication on all network interfaces by default, with no password required on fresh installations. Attackers can access the unauthenticated /api/tunnel/start endpoint to provision a public tunnel and republish the API at a publicly accessible address.

🎖@cveNotify

🚨 CVE-2026-86544
knowns versions before 0.30.0 contain an authorization bypass vulnerability where mutating code actions are incorrectly classified as read-only operations. Attackers with read-restricted sessions can exploit code.replace to modify permission configurations and escalate privileges on subsequent calls.

🎖@cveNotify

🚨 CVE-2026-66627
Unrestricted Upload of File with Dangerous Type vulnerability in EDGE22 Studios Ltd. GP Premium allows Remote Code Inclusion.

This issue affects GP Premium: from n/a through 2.5.5.

🎖@cveNotify

🚨 CVE-2026-32146
Improper path validation vulnerability in the Gleam compiler's handling of git dependencies allows arbitrary file system modification during dependency download.

Dependency names from gleam.toml and manifest.toml are incorporated into filesystem paths without sufficient validation or confinement to the intended dependency directory, allowing attacker-controlled paths (via relative traversal such as ../ or absolute paths) to target filesystem locations outside that directory. When resolving git dependencies (e.g. via gleam deps download), the computed path is used for filesystem operations including directory deletion and creation.

This vulnerability occurs during the dependency resolution and download phase, which is generally expected to be limited to fetching and preparing dependencies within a confined directory. A malicious direct or transitive git dependency can exploit this issue to delete and overwrite arbitrary directories outside the intended dependency directory, including attacker-chosen absolute paths, potentially causing data loss. In some environments, this may be further leveraged to achieve code execution, for example by overwriting git hooks or shell configuration files.

This issue affects Gleam from 1.9.0-rc1 until 1.15.4.

🎖@cveNotify

🚨 CVE-2026-32685
Path traversal vulnerability in Gleam's handling of custom documentation pages allows arbitrary file read and file write outside the intended documentation output directory.

The documentation.pages entries from gleam.toml are incorporated into filesystem paths without sufficient validation or confinement to the intended project and documentation output directories. The documentation.pages[].path field can be used to write generated documentation files outside the intended build/dev/docs/<package>/ output directory. The documentation.pages[].source field can be used to read files outside the project directory and embed their contents into generated documentation output.

An attacker who can convince a victim to run gleam docs build on an untrusted project, or with untrusted gleam.toml content, can cause local files readable by the victim to be included in generated documentation artifacts, and can cause generated documentation files to be written outside the intended docs output directory.

This issue affects Gleam from 1.16.0 until 1.17.0.

🎖@cveNotify

🚨 CVE-2026-42795
Symlink following vulnerability in Gleam's Hex package export allows files outside the project root to be embedded in the generated package tarball.

The file collection helpers (gleam_files, native_files, private_files) in compiler-cli/src/fs.rs use follow_links(true) when walking publishable directories such as src/ and priv/. The collected paths are added to the package archive via add_path_to_tar in compiler-cli/src/publish.rs without verifying that the resolved target remains within the project root. A symlink placed under a publishable directory will cause gleam export hex-tarball or gleam publish to embed the contents of the symlink target into the generated Hex package.

An attacker with write access to the project repository can place a symlink in src/ or priv/ pointing to an arbitrary file. When a maintainer or CI pipeline runs gleam publish or gleam export hex-tarball, local files readable by the publisher (such as secrets, tokens, or SSH keys) are silently embedded into the published package artifact.

This issue affects Gleam from 0.10.0-rc1 until 1.17.0.

🎖@cveNotify

🚨 CVE-2026-43965
Path traversal vulnerability in Gleam's dependency management allows arbitrary directory deletion via malicious build/packages/packages.toml content.

Package keys read from build/packages/packages.toml by LocalPackages::read_from_disc are passed without validation to paths.build_packages_package(), which constructs a filesystem path by joining the project build directory with the attacker-controlled key. The resulting path is then passed to fs::delete_directory (which calls remove_dir_all). No check is performed to ensure the path remains within the intended build/packages/ directory. Both absolute paths and relative traversal sequences (e.g. ../) are accepted as package keys, allowing deletion of arbitrary directories.

An attacker who can cause a victim to run gleam deps download on a project containing a malicious build/packages/packages.toml (e.g. by committing the normally-gitignored file to a repository) can cause arbitrary directories on the victim's system to be recursively deleted.

This issue affects Gleam from 0.18.0-rc1 until 1.17.0.

🎖@cveNotify

🚨 CVE-2026-48859
Observable Timing Discrepancy vulnerability in Erlang/OTP ssh (ssh_auth, ssh_options modules) allows unauthenticated remote username enumeration via timing side-channel in password authentication.

When the SSH daemon is configured with the user_passwords or password option, ssh_auth:check_password/3 performs a PBKDF2-SHA256 computation with 600,000 iterations (~300ms) for valid usernames, but returns immediately (~0ms) for invalid usernames via the ssh_options:get_password_option/2 path. This timing difference is detectable in a single authentication attempt and allows an unauthenticated attacker to distinguish valid from invalid usernames.

The user_passwords and password options are documented as intended for test purposes; the recommended alternative is pwdfun, which is not affected by this vulnerability.

This vulnerability is associated with program files lib/ssh/src/ssh_auth.erl and lib/ssh/src/ssh_options.erl.

This issue affects OTP from OTP 29.0 before OTP 29.0.2, corresponding to ssh from 6.0 before 6.0.1.

🎖@cveNotify

🚨 CVE-2026-54891
Improper Enforcement of Message Integrity During Transmission in a Communication Channel vulnerability in Erlang/OTP ssl (tls_gen_connection module) allows a network-positioned attacker to inject unauthenticated plaintext that the TLS client application later treats as authenticated server data.

The function tls_gen_connection:handle_protocol_record/3 rejects APPLICATION_DATA records that arrive in pre-handshake states when the TLS endpoint acts as a server, but does not apply the same check when the endpoint acts as a client. A network-positioned attacker can send plaintext APPLICATION_DATA records to the client during the handshake. The records are buffered and, once the handshake completes successfully, delivered to the application as if they were authenticated post-handshake data. The attacker cannot observe the client's response or steer the connection, so the impact is limited to blind injection of unauthenticated bytes. The injection window is wider for TLS versions prior to TLS 1.3 than for TLS 1.3.

This vulnerability is associated with program file lib/ssl/src/tls_gen_connection.erl.

TLS 1.3 is affected starting with OTP 22.0, when TLS 1.3 support was added.

This issue affects OTP from OTP 17.0 before OTP 27.3.4.14, from OTP 28.0 before OTP 28.5.0.3, and from OTP 29.0 before OTP 29.0.3, corresponding to ssl from 5.3.4 before 11.2.12.10, from 11.3 before 11.6.0.3, and from 11.7 before 11.7.3. Whether OTP before OTP 17.0, corresponding to ssl before 5.3.4, is affected is unknown.

🎖@cveNotify

🚨 CVE-2026-55953
The Erlang/OTP ssl TLS 1.2 (and earlier) and DTLS client does not verify that the cipher suite selected by the server in ServerHello was among the suites offered by the client in ClientHello. The client-side tls_handshake:hello/5 handler validates the negotiated protocol version and the downgrade sentinel but hands the server-chosen suite directly to ssl_handshake:handle_server_hello_extensions/9, which installs it without a membership check. The TLS 1.3 client path performs this check (per RFC 8446), so it is not affected.

An on-path attacker between the client and the intended server can respond with a ServerHello selecting an anonymous key exchange suite such as TLS_DH_anon_* or TLS_ECDH_anon_* that the client never offered. Anonymous suites do not require the server to present a certificate, so the entire verify_peer and cacerts configuration is bypassed: the attacker completes the handshake with its own ephemeral parameters, no certificate is validated, no hostname is checked, and ssl:connect returns {ok, Socket}. All subsequent application traffic is readable and modifiable by the attacker.

This issue affects OTP from OTP 17.0 before OTP 27.3.4.15, from OTP 28.0 before OTP 28.5.0.4, and from OTP 29.0 before OTP 29.0.4, corresponding to ssl from 5.3.4 before 11.2.12.11, from 11.3 before 11.6.0.4, and from 11.7 before 11.7.4. Whether OTP before OTP 17.0, corresponding to ssl before 5.3.4, is affected is unknown.

🎖@cveNotify

🚨 CVE-2026-59250
Classic buffer overflow in the Erlang/OTP megaco flex scanner C driver allows a remote unauthenticated attacker to corrupt the driver's memory (and potentially achieve remote code execution or a denial-of-service crash) by sending a single text-encoded H.248/Megaco message containing an oversized property parm name.

When tokenizing a Local/Remote descriptor, mfs_load_property_groups extracts the attacker-controlled property name (bounded only by the message length) and, when no value follows, formats it into a fixed 512-byte error_msg field of the MfsErlDrvData struct using an unchecked sprintf call. Names longer than roughly 452 bytes overflow into the immediately following struct fields (text_buf, text_ptr, term_spec, term_spec_size, term_spec_index), overwriting live pointers and counters with attacker-chosen bytes. Subsequent scanner code writes and frees through the corrupted pointers, producing arbitrary write and arbitrary free primitives inside the BEAM VM process, which can be leveraged for remote code execution. On builds compiled with _FORTIFY_SOURCE the overflow is detected at runtime and terminates the process with SIGABRT, resulting in denial of service.

The overflow occurs in the flex scanner before any grammar or Megaco-level authentication processing, so exploitation requires only network reachability to the megaco transport port on a node configured with {scanner, flex}.

This vulnerability is associated with program files lib/megaco/src/flex/megaco_flex_scanner_drv.flex.src and program routines mfs_load_property_groups.

This issue affects OTP from OTP 17.0 before OTP 27.3.4.15, from OTP 28.0 before OTP 28.5.0.4, and from OTP 29.0 before OTP 29.0.4, corresponding to megaco from 3.17.1 before 4.7.2.2, from 4.8 before 4.8.3.1, and from 4.9 before 4.9.1. Whether OTP before OTP 17.0, corresponding to megaco before 3.17.1, is affected is unknown.

🎖@cveNotify