Kotaemon through 0.12.0 fails to properly validate conversation ownership in select_conv, delete_conv, rename_conv, and on_set_public_conversation functions in control.py. Attackers can read other users' chat histories, delete conversations, or rename conversations by supplying arbitrary conversation identifiers without proper authorization checks.

Quivr through 0.0.322 fails to validate ownership in prompt endpoints, allowing authenticated users to modify any prompt by identifier. Attackers with read-only access to shared brains can read exposed prompt identifiers and overwrite system prompts affecting all brain users.

HyperDX through 1.10.1 fails to enforce role-based access controls in team management endpoints, allowing any team member to perform administrative actions. Attackers can delete team members including owners, rotate API keys, and rename teams by sending requests to PATCH /team/apiKey, PATCH /team/name, and DELETE /team/member endpoints.

BISHENG before 2.6.0 contains a remote code execution vulnerability in the workflow run_once endpoint that allows authenticated users to execute arbitrary Python code. Attackers can submit crafted Code node definitions to the POST /api/v1/workflow/run_once endpoint, which executes them with exec() without sandboxing, gaining access to filesystem, credentials, and internal network resources.

Argo Rollouts dashboard through 1.10.0 binds to all interfaces and exposes mutating Rollout operations without authentication, authorization, or CSRF protection. Attackers on the same network can invoke PromoteRollout, AbortRollout, RestartRollout, SetRolloutImage, UndoRollout, and RetryRollout operations across all namespaces accessible to the operator's kubeconfig.

StarRocks through 4.0.13 contains an authentication bypass vulnerability in five REST handler classes that override execute() directly instead of implementing executeWithoutPassword(). Attackers can access six unauthenticated endpoints on the frontend HTTP port to disclose cluster topology, database metadata, JVM statistics, and version information without credentials.

Qwen-Agent through 0.0.34 contains a path traversal vulnerability in the document parser that fails to restrict file access to intended directories. Attackers can supply absolute file paths to the unauthenticated Gradio interface to read arbitrary files accessible by the server process.

Twenty through 2.35.0 contains an open redirect vulnerability in the OAuthPropagatorController.propagateOAuthCallback endpoint that treats the state query parameter as a redirect URL. Attackers can craft malicious requests to redirect users to arbitrary hosts while forwarding OAuth authorization codes, bypassing domain validation when IS_MULTIWORKSPACE_ENABLED is disabled.

Mastra through 1.63.0 contains an authentication bypass vulnerability in the memory API thread ownership validation when mapUserToResourceId callback is omitted from configuration. Authenticated attackers can enumerate all threads via GET /api/memory/threads and read conversation history and metadata of other resource owners.

Immich through 3.1.0 fails to properly enforce locked asset visibility when assets are locked through the single-asset endpoint, allowing them to remain accessible through shared albums and links. Attackers can read locked assets and their metadata by accessing existing shared albums or links, bypassing the locked visibility protection.

R2R through 3.6.5 fails to properly validate user ownership in conversation update and message handlers, allowing authenticated users to modify other users' conversations. Attackers can supply arbitrary conversation identifiers to rename conversations and append messages to other users' conversation histories, corrupting state and injecting malicious content.

Portkey AI Gateway through 1.15.2 contains a server-side request forgery vulnerability in the /v1/proxy/* route that lacks requestValidator middleware. Attackers can set the x-portkey-custom-host header to internal addresses and forward requests with Authorization headers to reach internal services and exfiltrate provider API keys.

Gophish through 0.12.1 fails to enforce account lockout and password change requirements in the API authentication middleware. Attackers with valid API keys can bypass these security controls and retain full API access even when their account is locked or password change is required.

Qwen-Agent through 0.0.34 contains a server-side request forgery vulnerability in the document parsing path that treats caller-supplied paths as URLs without scheme restriction or host validation. Attackers can reach the unauthenticated Gradio interface to make the server issue HTTP requests to arbitrary internal addresses including metadata services and read retrieved content through parsed document output.

Komodo through 2.3.2 discloses internal resource identifiers and writes audit entries before performing permission checks in the /execute and /execute/{variant} handlers. Authenticated users can guess resource names to obtain internal identifiers and insert fraudulent audit log entries misrepresenting privileged operations.

Redpanda through 26.2.2 binds the Admin API to 0.0.0.0:9644 with admin_api_require_auth defaulting to false, treating unauthenticated requests as superusers. Attackers can reach port 9644 without credentials to create and delete broker accounts, modify cluster configuration, and disrupt partition replication.

Zipkin through 3.6.1 exposes Spring Boot Actuator endpoints on the tracing API port without authentication, allowing unauthenticated attackers to access sensitive information. Attackers can read environment variables, bean configurations, and storage credentials via actuator endpoints, or modify log levels to suppress logging.

Duplicacy through 3.2.5 contains a path traversal vulnerability in the restore function that fails to validate entry paths deserialized from snapshot files. Attackers can craft malicious snapshot entries with directory traversal sequences to write files outside the restore directory to arbitrary locations accessible by the restoring user.

Logto through 1.42.0 contains a server-side request forgery vulnerability in the OIDC SSO connector creation endpoint that fails to validate the issuer URL parameter. Tenant administrators with Management API credentials can supply arbitrary internal URLs to trigger HTTP GET requests to private network services, with response content returned in API responses.

Logto through 1.42.0 contains a server-side request forgery vulnerability in the POST /api/hooks/:id/test endpoint that accepts arbitrary URLs without host validation. Tenant administrators with Management API tokens can make the server issue HTTP POST requests to internal URLs and retrieve response bodies from services on the private network.

Hermes Agent 0.18.2 prior to 0.19.0 contains a supply chain vulnerability in its bundled MCP catalog that allows a remote attacker to execute arbitrary code by compromising a third-party upstream repository referenced via a mutable branch rather than a pinned commit SHA. An attacker who compromises the upstream repository can propagate malicious code to every host that installs the affected catalog entry, with no further action required by the operator.

Hermes Agent 0.16.0 prior to 0.17.0 contains an improper path restriction vulnerability that allows attackers who can influence ingested message content to overwrite the credential store by bypassing sensitive-path guards that excluded the auth.json file. Attackers can craft malicious messages directing the agent's file-write tooling to overwrite the credential store without triggering any path-based protection, enabling credential tampering or unauthorized access.

Improper limitation of a pathname to a restricted directory in the aws:downloadContent plugin in amazon-ssm-agent before 3.3.4515.0 might allow an authenticated remote user whose ssm:SendCommand permission is restricted to the AWS-DownloadContent document, to write arbitrary files outside the intended download directory with root privileges, via crafted object keys in the S3 source the document is directed to retrieve. This issue may lead to arbitrary code execution as root if specific sensitive files are overwritten. To remediate this issue, customers should upgrade amazon-ssm-agent to version 3.3.4515.0 or later.

Flextype CMS through v1.0.0-dev contains an expression language injection vulnerability that allows authenticated attackers with a valid API token to read arbitrary files by passing unsanitized user-supplied input to the Symfony ExpressionLanguage engine via the POST /api/v1/query endpoint. Attackers can leverage exposed application objects including filesystem() and serializers() within the evaluation scope to read arbitrary server files and achieve conditional remote code execution if a PHP file can be placed on disk through a secondary vector.

In MongoDB Connector for BI, MongoDB object names such as collection, field, and index names are placed into the quoted identifiers of the DDL text returned by SHOW CREATE statements without escaping the identifier delimiter. A user with permission to write to a sampled MongoDB collection can choose a name that closes the quoted identifier early, so that additional SQL text becomes part of the generated output. If an operator or automated tool later replays that generated statement against a SQL server, the additional text is executed with the privileges of that session.