CVE Feed

Coder allows organizations to provision remote development environments via Terraform. Starting in version 2.24.0 and prior to versions 2.29.7, 2.32.7, 2.33.8, and 2.34.2, `NewDataBuilder` in `provisionersdk/proto/dataupload.go` allocated a byte slice using the client-supplied `FileSize` from a `DataUpload` message without an upper-bound check. Although the DRPC wire limit is 4 MiB, the `FileSize` value itself was unconstrained. The fix in versions 2.29.7, 2.32.7, 2.33.8, and 2.34.2 validates `FileSize` against an upper bound (`MaxFileSize = 100 MiB`) before allocation. As a workaround, restrict access to the provisioner daemon serve endpoint to trusted provisioner daemon service accounts.

Coder allows organizations to provision remote development environments via Terraform. Starting in version 2.17.0 and prior to versions 2.29.7, 2.32.7, 2.33.8, and 2.34.2, `POST /api/v2/files` converts zip uploads to tar in memory via `CreateTarFromZip`, which enforced a per-entry size limit but no aggregate limit on total decompressed output, writing to an unbounded in-memory buffer. Exploitation requires authenticated file-upload access and the impact is limited to availability (denial of service). The fix in versions 2.29.7, 2.32.7, 2.33.8, and 2.34.2 adds a metadata preflight check that sums projected entry sizes and a streaming writer that enforces the aggregate limit during decompression. As a workaround, restrict file-upload permissions to trusted users or place a reverse proxy with request-body size limits in front of `coderd`.

Coder allows organizations to provision remote development environments via Terraform. Prior to versions 2.29.7, 2.32.7, 2.33.8, and 2.34.2, the `PUT /api/v2/users/{user}/password` endpoint authorized only `ActionUpdatePersonal` and did not prevent a `user-admin` from resetting an `owner` account's password. It also did not require the current password when an admin reset another user's password. Exploitation requires the privileged `user-admin` role so practical risk is limited to deployments that grant `user-admin` to less trusted operators. The fix in versions 2.29.7, 2.32.7, 2.33.8, and 2.34.2 prevents non-owner users from resetting the password of an account that holds the `owner` role. As a workaround, restrict the `user-admin` role to trusted administrators.

Coder allows organizations to provision remote development environments via Terraform. Prior to versions 2.29.7, 2.32.7, 2.33.8, and 2.34.2, Coder's OIDC callback checked `email_verified` with a direct Go `bool` type assertion. When an IdP returned the claim as a non-boolean (for example the string `"false"`) or omitted it, the assertion failed open and the email was treated as verified. Combined with an unconditional email-based account fallback, this enabled account takeover. The fix in versions 2.29.7, 2.32.7, 2.33.8, and 2.34.2 coerces `email_verified` across bool, string and numeric types (fail-closed) and blocks the email fallback when the matched user already has a different linked IdP subject. As a workaround, ensure the IdP returns `email_verified` as a native JSON boolean. The email-fallback linking issue has no configuration workaround; upgrading is required.

Cap's GET /api/video/ai endpoint fails to validate user ownership or membership before returning private video AI metadata including titles, summaries, and chapters. Authenticated attackers can supply arbitrary video IDs to read sensitive AI-generated content and trigger unauthorized AI generation that consumes the video owner's credits without consent.

String::Util versions before 1.36 for Perl are susceptible to a regular expression denial of service. The trim and rtrim functions stripped trailing whitespace with s/\s*$//u. Because \s* matches greedily and the $ anchor fails whenever a non-whitespace character follows the whitespace, the regex engine retries the match at each offset of a long whitespace run, producing quadratic backtracking. The fix replaces \s*$ with \s+$. Any caller that passes untrusted input to trim or rtrim can trigger CPU exhaustion with a string containing a long run of whitespace.

mem0's openmemory/api component contains an unauthenticated access vulnerability that allows unauthenticated attackers to read, write, and delete arbitrary user memories by accessing API routers registered without authentication middleware. Attackers can supply arbitrary user_id parameters or directly access memory retrieval endpoints to expose private memory content, or invoke pause endpoints with global_pause=true to cause denial-of-service across all users.

DBI versions before 1.650 for Perl read one byte out-of-bounds in preparse when deleting an initial SQL comment. The preparse method normalises SQL and removes comments. When the SQL starts with a comment line, the deletion of that line during normalisation led to an out-of-bounds read by one byte. The result is a fault on memory-hardened builds and nondeterministic newline retention on normal builds.

DBI versions before 1.650 for Perl have a heap overflow when preparsing SQL statements with an extreme number of placeholders. The fix for CVE-2026-10879 did not allocate enough memory to handle approximately 1.2-million placeholders. DBI version 1.650 sets a hard limit of 99,999 placeholders.

DBI versions before 1.650 for Perl are vulnerable to code injection via caller-influenced Profile. When a string is assigned to a DBI handle's Profile attribute, DBI splits it into path, package and arguments, and interpolates the package part in a string eval with no validation of the package name. Any caller-influenced value that reaches the Profile attribute is therefore arbitrary Perl code execution, including calls to run system commands. The Profile attribute can be set from three different sources that can carry untrusted data: the DBI_PROFILE environment variable, a direct attribute assignment, and a DSN driver-attribute clause dbi:Driver(Profile=>SPEC):db. An attacker controlling any of those inputs runs arbitrary Perl in the host process. The strongest remote position is a network-exposed DBI::Gofer / DBI::ProxyServer whose per-request DSN reaches the Profile attribute, letting a client execute code on the broker host.

The application contains a stack-based buffer overflow vulnerability that can be exploited by an attacker to execute arbitrary code.

The application contains a use-after-free vulnerability that can be exploited to cause memory corruption while parsing specially crafted files. This could allow an attacker to execute arbitrary code in the context of the current process.

The application contains an out-of-bounds write vulnerability that can be exploited by an attacker to cause the program to write data past the end of an allocated memory buffer. This can lead to arbitrary code execution.

Coder allows organizations to provision remote development environments via Terraform. Prior to versions 2.29.7, 2.32.7, 2.33.8, and 2.34.2, two flaws in Coder's OIDC login chained into account takeover. Email-based user matching fell back to linking by email without checking for an existing link to a different IdP subject and the `email_verified` claim was only enforced when present as a boolean `false` so an absent or non-boolean claim was treated as verified. The fix in versions 2.29.7, 2.32.7, 2.33.8, and 2.34.2 restricts the email fallback to first-time and legacy linking and defaults `email_verified` to false when the claim is absent or of an unexpected type. As a workaround, configure the OIDC provider to disallow self-registration or to require email verification before issuing tokens.

Hasura is an open-source product that provides users GraphQL or REST APIs. Prior to 2.49.2 and 2.45.5, a user can use a where clause on a table computed field (returning SETOF some_table) to infer row values that ought to be filtered for their role based on some_table's row-level permissions. While such rows cannot be returned directly, like predicates on strings for instance allow values to be brute forced efficiently with the where clause as an oracle. This issue is fixed in versions 2.49.2 and 2.45.5.

FastGPT is a knowledge-based AI application platform. Prior to 4.15.0, GET /api/core/ai/record/getRecord authenticates the caller but loads LLM request and response traces only by requestId without team scoping, allowing any authenticated user to read another team's prompts, retrieved RAG chunks, and completions if the requestId is known. This issue is fixed in version 4.15.0.

FastGPT is a knowledge-based AI application platform. Prior to 4.15.0-beta4, the HTTP-tool OpenAPI schema importer validates only the top-level URL before passing it to SwaggerParser.bundle, whose remote reference resolver fetches $ref URLs without FastGPT's internal-address guard and returns fetched content inline, allowing an authenticated team member to read internal services or cloud metadata. This issue is fixed in version 4.15.0-beta4.

FastGPT is an open source AI knowledge base platform. Prior to v4.15.0-beta5, two FastGPT file handlers authorize an unrelated resource and then sign or read an S3 object using a key taken directly from the request, without checking that the key belongs to the caller's team. Because S3 object keys are global within the bucket and carry the tenant id only as a path segment, an attacker can supply another team's key and obtain its file contents through the chat-file presign endpoint or dataset preview endpoint. This issue is fixed in version v4.15.0-beta5.

FastGPT is an open source AI knowledge base platform. From 4.14.17 to before 4.15.0-beta4, FastGPT allows an authenticated tenant user to call POST /api/core/dataset/collection/create/reTrainingCollection in a way that persists a server-owned datasetId value from another tenant. This creates mixed dataset objects and downstream dataset, collection, and training endpoints then make authorization decisions from inconsistent ownership anchors, allowing cross-tenant read, update, and delete access when mixed object ids are known. This issue is fixed in version 4.15.0-beta4.

Anki is a program for creating and reviewing flashcards. Prior to 25.09.4, Anki's webview-based pages communicate with the Rust backend using an internal localhost API, and user scripts included via iframes in the editor can access this API despite protections intended to block reviewer and editor scripts. A malicious imported card package with an embedded iframe can use exposed API methods such as getImageForOcclusion to read arbitrary files accessible to the Anki process and exfiltrate them over the network. This issue is fixed in version 25.09.4.

Anki is a program for creating and reviewing flashcards. Prior to 25.09.3, Anki launches a local HTTP server to serve media files and web pages for parts of its interface, but requests from other origins were not sufficiently blocked. A malicious website could potentially trigger side-effecting requests to the local server, with severity varying by browser depending on Private Network Access protections. This issue is fixed in version 25.09.3.

Coder allows organizations to provision remote development environments via Terraform. In versions prior tp 2.24.5, 2.29.13, 2.30.8, 2.31.12, 2.32.2, and 2.33.3, `azureidentity.Validate()` verifies that the PKCS#7 signer certificate chains to a trusted Azure CA but never verifies the PKCS#7 signature itself. An attacker can embed a legitimate Azure certificate alongside arbitrary content e.g. `{"vmId":"<target>"}` and the forged `vmId` will be accepted returning the victim workspace agent's session token. No authentication is required. The attacker only needs to know a target VM's `vmId` which is a `UUIDv4`. That's a practical limitation which would typically require prior access to be exploited. Versions 2.24.5, 2.29.13, 2.30.8, 2.31.12, 2.32.2, and 2.33.3 patch the issue. As a workaround, reconfigure any Azure templates to use token authentication rather than `azure-instance-identity`.

OpenWrt is a Linux operating system targeting embedded devices. Before v25.12.5, an integer underflow in handle_send_a() of the Emergency Access Daemon allows any unauthenticated attacker on the local network to crash the daemon by sending a single crafted UDP packet. The message length underflows before a bounds check and is then passed to memcpy as a very large size. This issue is fixed v25.12.5.

The public dashboard deletion endpoint does not enforce organization isolation, allowing an Org Admin in one organization to delete public dashboards belonging to a different organization by supplying the target dashboard's identifiers.

Coder allows organizations to provision remote development environments via Terraform. Versions prior tp 2.24.5, 2.29.13, 2.30.8, 2.31.12, 2.32.2, and 2.33.3 are vulnerable to unauthenticated semi-blind Server-Side Request Forgery (SSRF) via the Azure instance identity endpoint (`POST /api/v2/workspaceagents/azure-instance-identity`). An external attacker can force the Coder server to issue HTTP GET requests to arbitrary internal or external hosts by submitting a crafted PKCS#7 signature. The server does not return the target's response body, but error messages in the API response reveal whether the target is reachable and what type of failure occurred. Versions 2.24.5, 2.29.13, 2.30.8, 2.31.12, 2.32.2, and 2.33.3 patch the issue. As a workaround, if the Azure identity-auth mechanism is not being used then restrict access to the corresponding endpoint (`/api/v2/workspaceagents/azure-instance-identity`) using ingress firewall and/or proxy ACLs.

Koodo Reader is an ebook reader. In version 2.3.0 and earlier, Koodo Reader is vulnerable to remote code execution through malicious EPUB files because the open-book IPC handler enables nodeIntegrationInSubFrames and EPUB chapter content is rendered with unsanitized innerHTML. An attacker can craft an EPUB book that, when imported and opened by the victim, instantiates a hidden iframe with Node.js API access and executes arbitrary operating system commands with the victim user's privileges. This issue is fixed in version 2.3.1.

mem0 contains unauthenticated config API endpoints that expose LLM API keys in plaintext and allow server-side request forgery via attacker-controlled ollama_base_url parameter. Unauthenticated attackers can retrieve stored secrets like OpenAI API keys via GET /api/v1/config/ or trigger SSRF attacks by setting ollama_base_url to internal addresses like cloud IMDS via PUT /api/v1/config/mem0/llm endpoint.

Actual is a local-first personal finance app. Prior to 26.6.0, in OpenID multi-user mode, disabling a user only blocks future OpenID login for that identity, while existing Actual session tokens for the disabled user remain valid. The shared session validation path accepts any existing token row that has not expired without checking whether the associated user is still enabled, allowing a disabled user to continue calling authenticated server endpoints. This issue is fixed in version 26.6.0.

Actual is a local-first personal finance tool. Prior to 26.6.0, exportToCSV and exportQueryToCSV in packages/loot-core/src/server/transactions/export/export-to-csv.ts pass user-controlled Payee, Notes, Account, and Category strings to csv-stringify with no cast callback and no formula-prefix neutralization. Strings that begin with equals sign, plus, minus, at sign, tab, or carriage return survive verbatim into the exported CSV, and when a recipient opens the file in Excel, LibreOffice Calc, or Google Sheets, the strings are interpreted as formulas, enabling transaction data exfiltration and attacker-chosen spreadsheet display values. This issue is fixed in version 26.6.0.

Actual is a local-first personal finance tool. Prior to 26.6.0, the GET /secret/:name endpoint in @actual-app/sync-server checks only that the caller has a valid session and does not verify the caller is an admin, while the sibling POST /secret/ handler enforces an admin check in OpenID mode. Any authenticated non-admin BASIC user in OpenID multi-user deployments can probe the secrets store and learn which admin-managed bank-sync integrations have been configured, including simplefin_accessKey, pluggyai_clientSecret, pluggyai_itemIds, and the gocardless secrets. This issue is fixed in version 26.6.0.

Actual is a local-first personal finance app. Prior to 26.6.0, @actual-app/cli ships a hand-rolled CSV serializer in packages/cli/src/output.ts used whenever the global --format csv option is passed, whose escapeCsv helper only handles RFC 4180 delimiter, quote, and newline escaping and does not neutralize standard CSV formula-injection prefixes. Any CLI command that streams an object array containing user-controlled strings, including transactions list, accounts list, payees list, categories list, tags list, category-groups list, rules list, schedules list, and query, can emit cells that auto-evaluate when the resulting CSV is opened in Excel, LibreOffice Calc, or Google Sheets, enabling data exfiltration and arbitrary formula execution. This issue is fixed in version 26.6.0.

Actual is an open-source personal finance application. Prior to 26.7.0, a missing authorization issue allows a shared user with user_access on a budget file to perform owner-only file management actions. A non-owner shared user can call file-management endpoints intended for higher-privilege users, including /delete-user-file, /reset-user-file, and /user-create-key, because requireFileAccess treats ordinary shared access as sufficient for file-management operations that should be restricted to the file owner or an administrator. This issue is fixed in version 26.7.0.

Serena is a powerful MCP toolkit for coding that provides semantic retrieval and editing capabilities. Prior to v1.5.2, Serena's built-in web dashboard exposes an unauthenticated Flask API on a fixed, predictable port, with no authentication, no CSRF protection, and no Host header validation. A DNS rebinding attack allows a malicious webpage to reach this API from any browser and write arbitrary content to the agent's persistent memory store, which the agent reads and acts on autonomously. Combined with execute_shell_command using shell=True, this creates a remote code execution chain requiring only that the victim visit a malicious webpage while Serena is running. This issue is fixed in version v1.5.2.

Dashy is a self-hostable personal dashboard. Prior to 4.3.7, Dashy's workspace view trusts the url query parameter and assigns it directly to an iframe source without scheme validation. If a logged-in user opens a crafted workspace link containing a javascript: URL, JavaScript runs on the Dashy origin and can read same-origin browser data, interact with the Dashy DOM, and send requests as the victim. This issue is fixed in version 4.3.7.

calibre is an e-book manager. Prior to 9.10.0, a malicious EPUB, OPF, or PDF file can execute arbitrary Python code when its metadata is read by calibre, including through Add books or Edit books, by embedding a custom column definition with a python: template in calibre:user_metadata that is passed unsanitized to exec() in the template formatter. This issue is fixed in version 9.10.0.

Cilium is a networking, observability, and security solution. Prior to 1.17.16, from 1.18.2 to 1.18.9, and from 1.19.0 to 1.19.3, users with the ability to create CiliumLocalRedirectPolicies can specify arbitrary ClusterIPs via addressMatcher, enabling hijacking traffic to Services in any namespace and bypassing namespace scoping enforced by serviceMatcher; deleting such a policy can also corrupt Cilium internal service state and stop service translation for the affected Service. This issue is fixed in versions 1.17.16, 1.18.10, and 1.19.4.

DataEase is an open source data visualization and analysis tool. Prior to 2.10.24, a share mode chart data interface only validates that sceneId matches the resourceId in the link token and fails to validate whether tableId and field IDs in the request body belong to the shared resource, allowing an attacker with a valid share link token to replace dataset identifiers and retrieve unauthorized data through POST /de2api/chartData/getData. This issue is fixed in version 2.10.24.

DataEase is an open source data visualization and analysis tool. Prior to 2.10.24, the /de2api/share/proxyInfo share interface generates and returns X-DE-LINK-TOKEN before validating the share password or ticket, allowing unauthenticated attackers who know a protected share UUID to obtain a valid link token for subsequent share-related API calls even with missing or invalid credentials. This issue is fixed in version 2.10.24.

DataEase is an open source data visualization and analysis tool. Prior to 2.10.24, dashboard text components render stored component content with Vue v-html without server-side HTML sanitization, allowing an authenticated user who can edit dashboard component data to inject HTML with executable event handlers that execute when another user or shared-link visitor views the dashboard. This issue is fixed in version 2.10.24.

LocalAI contains an unauthenticated server-side request forgery vulnerability in the POST /models/apply endpoint that allows attackers to fetch arbitrary internal URLs. The endpoint passes unsanitized gallery URL fields directly to gallery.GetGalleryConfigFromURLWithContext without proper validation, enabling attackers to force the server to issue HTTP GET requests to private and loopback ranges with partial response content leaked through error messages.

DataEase is an open source data visualization and analysis tool. Prior to 2.10.24, chart quota and Y-axis filters embed attacker-controlled filter values directly into generated SQL in Quota2SQLObj.getYWheres() without applying the SQL literal validation and escaping used by other filter paths, allowing an authenticated user who can create or modify chart definitions or submit chart data requests containing quota filters to inject SQL into queries executed against configured datasources. This issue is fixed in version 2.10.24.

Cognee before 1.2.0 contains an improper access control vulnerability that allows unauthenticated attackers to overwrite the global LLM provider configuration by self-registering an account and calling the settings endpoint, which performs no admin or superuser check. Attackers can redirect all LLM operations instance-wide to an attacker-controlled endpoint by exploiting the process-wide singleton configuration cache, enabling exfiltration of prompts, uploaded documents, extracted entities, and knowledge graph content from all users.

FluxInk (formerly Sunia SPB Peripheral) Color Management Driver (TcnPeripheral64.sys) 1.0.7.2 allows local privilege escalation for a standard user account via arbitrary physical memory mapping at \Device\PhysicalMemory. Fixed in version 1.0.7.6. The fixed driver is currently available in the Windows 11 25H2 HLK (Hardware Lab Kit). The fixed driver may be available through Windows Update or from Lenovo directly.

DataEase is an open source data visualization and analysis tool. Prior to 2.10.24, ShareSecretManage uses a hardcoded default share link signature key, allowing an attacker who can obtain a passwordless share for a resource and user to use the known key link-pwd-fit2cloud to forge linkToken JWTs, bypass TokenFilter verification, and access backend resources as the share creator even if the original share has been revoked. This issue is fixed in version 2.10.24.

DataEase is an open source data visualization and analysis tool. Prior to 2.10.24, the H2 database JDBC URL validation logic can be bypassed with special Unicode characters whose case-conversion behavior differs between DataEase validation and H2 parsing, allowing attackers to smuggle dangerous parameters such as init in malicious H2 JDBC connection strings and achieve arbitrary code execution. This issue is fixed in version 2.10.24.

DataEase is an open source data visualization and analysis tool. Prior to 2.10.24, the /de2api/datasetData/previewSql endpoint lacks the mandatory @DePermit permission validation annotation, allowing any authenticated user to specify datasourceId=-1, access the built-in engine database, execute arbitrary SQL statements, and read sensitive core data. This issue is fixed in version 2.10.24.

DataEase is an open source data visualization and analysis tool. Prior to 2.10.24, a bypass of the H2 zip protocol and file dropper fix allows an authenticated attacker to upload a zip archive disguised with a .ttf extension through FontManage.saveFile and then exploit it through the zip protocol to achieve remote code execution. This issue is fixed in version 2.10.24.

DataEase is an open source data visualization and analysis tool. Prior to 2.10.24, the font management module allows authenticated users to submit an arbitrary fileTransName when creating a font record; when the record is later deleted, the backend concatenates that stored value with the font storage directory and passes it to FileUtils.deleteFile() without path traversal sanitization, allowing deletion of arbitrary writable files in the application container. This issue is fixed in version 2.10.24.

DataEase is an open source data visualization and analysis tool. Prior to 2.10.24, any authenticated user can download (/exportCenter/download/{id}), delete (/exportCenter/delete), retry (/exportCenter/retry/{id}), or generate download links (/exportCenter/generateDownloadUri/{id}) for export tasks belonging to other users by manipulating the task ID parameter, and the /exportCenter/download/{id} endpoint is whitelisted from authentication, allowing unauthenticated access to exported files. This issue is fixed in version 2.10.24.

Coder allows organizations to provision remote development environments via Terraform. Prior to versions 2.29.7 and 2.30.2, the `dotfiles` registry module passed unsanitized user input to shell commands, allowing arbitrary code execution inside a provisioned workspace. Any user who supplied a crafted `dotfiles_uri` value (for example, one containing shell command substitution such as `$(...)`) could achieve command execution in their own workspace. The Create Workspace page's `mode=auto` deep links amplified this into a one-click attack: an attacker could craft a URL that prefilled `param.dotfiles_uri` and silently provisioned a workspace with the attacker-controlled value, with no explicit user confirmation. In versions 2.29.7 and 2.30.2, input validation was added to the dotfiles module to reject URIs and usernames containing special characters, and the unsafe `eval`/`sh -c` usage was removed. This eliminated the command injection at its source.