PraisonAI: Sandbox Escape via shell=True and Bypassable Blocklist in SubprocessSandbox
PraisonAI is a multi-agent teams system. Prior to version 4.5.97, SubprocessSandbox in all modes (BASIC, STRICT, NETWORK_ISOLATED) calls subprocess.run() with shell=True and relies solely on string-pattern matching to block dangerous commands. The blocklist does not include sh or bash as standalone executables, allowing trivial sandbox escape in STRICT mode via sh -c '<command>'. This issue has been patched in version 4.5.97.
PraisonAI: SSRF in FileTools.download_file() via Unvalidated URL
PraisonAI is a multi-agent teams system. Prior to version 1.5.95, FileTools.download_file() in praisonaiagents validates the destination path but performs no validation on the url parameter, passing it directly to httpx.stream() with follow_redirects=True. An attacker who controls the URL can reach any host accessible from the server including cloud metadata services and internal network services. This issue has been patched in version 1.5.95.
PraisonAI: Authentication Bypass in OAuthManager.validate_token()
PraisonAI is a multi-agent teams system. Prior to version 4.5.97, OAuthManager.validate_token() returns True for any token not found in its internal store, which is empty by default. Any HTTP request to the MCP server with an arbitrary Bearer token is treated as authenticated, granting full access to all registered tools and agent capabilities. This issue has been patched in version 4.5.97.
PraisonAI: Missing Authentication in WebSocket Gateway
PraisonAI is a multi-agent teams system. Prior to version 4.5.97, the PraisonAI Gateway server accepts WebSocket connections at /ws and serves agent topology at /info with no authentication. Any network client can connect, enumerate registered agents, and send arbitrary messages to agents and their tool sets. This issue has been patched in version 4.5.97.
PraisonAI: ReDoS via Unvalidated User-Controlled Regex in MCPToolIndex.search_tools()
PraisonAI is a multi-agent teams system. Prior to version 4.5.90, MCPToolIndex.search_tools() compiles a caller-supplied string directly as a Python regular expression with no validation, sanitization, or timeout. A crafted regex causes catastrophic backtracking in the re engine, blocking the Python thread for hundreds of seconds and causing a complete service outage. This issue has been patched in version 4.5.90.
PraisonAI: Python Sandbox Escape via str Subclass startswith() Override in execute_code
PraisonAI is a multi-agent teams system. Prior to version 1.5.90, execute_code() in praisonai-agents runs attacker-controlled Python inside a three-layer sandbox that can be fully bypassed by passing a str subclass with an overridden startswith() method to the _safe_getattr wrapper, achieving arbitrary OS command execution on the host. This issue has been patched in version 1.5.90.
PraisonAI: Shell Injection in run_python() via Unescaped $() Substitution
PraisonAI is a multi-agent teams system. Prior to version 1.5.90, run_python() in praisonai constructs a shell command string by interpolating user-controlled code into python3 -c "<code>" and passing it to subprocess.run(..., shell=True). The escaping logic only handles \ and ", leaving $() and backtick substitutions unescaped, allowing arbitrary OS command execution before Python is invoked. This issue has been patched in version 1.5.90.
PraisonAI: SSRF via Unvalidated api_base in passthrough() Fallback
PraisonAI is a multi-agent teams system. Prior to version 4.5.90, passthrough() and apassthrough() in praisonai accept a caller-controlled api_base parameter that is concatenated with endpoint and passed directly to httpx.Client.request() when the litellm primary path raises AttributeError. No URL scheme validation, private IP filtering, or domain allowlist is applied, allowing requests to any host reachable from the server. This issue has been patched in version 4.5.90.
PraisonAI: Second-Order SQL Injection in `get_all_user_threads`
PraisonAI is a multi-agent teams system. Prior to version 4.5.90, the get_all_user_threads function constructs raw SQL queries using f-strings with unescaped thread IDs fetched from the database. An attacker stores a malicious thread ID via update_thread. When the application loads the thread list, the injected payload executes and grants full database access. This issue has been patched in version 4.5.90.
PraisonAI: OS Command Injection in MCPHandler.parse_mcp_command()
PraisonAI is a multi-agent teams system. From version 4.5.15 to before version 4.5.69, the --mcp CLI argument is passed directly to shlex.split() and forwarded through the call chain to anyio.open_process() with no validation, allowlist check, or sanitization at any hop, allowing arbitrary OS command execution as the process user. This issue has been patched in version 4.5.69.
Mesop: Unbounded Thread Creation in WebSocket Handler Leads to Denial of Service
Mesop is a Python-based UI framework that allows users to build web applications. From version 1.2.3 to before version 1.2.5, an uncontrolled resource consumption vulnerability exists in the WebSocket implementation of the Mesop framework. An unauthenticated attacker can send a rapid succession of WebSocket messages, forcing the server to spawn an unbounded number of operating system threads. This leads to thread exhaustion and Out of Memory (OOM) errors, causing a complete Denial of Service (DoS) for any application built on the framework. This issue has been patched in version 1.2.5.
LTI JupyterHub Authenticator: Unbounded Memory Growth via Nonce Storage (Denial of Service)
LTI JupyterHub Authenticator is a JupyterHub authenticator for LTI. Prior to version 1.6.3, the LTI 1.1 validator stores OAuth nonces in a class-level dictionary that grows without bounds. Nonces are added before signature validation, so an attacker with knowledge of a valid consumer key can send repeated requests with unique nonces to gradually exhaust server memory, causing a denial of service. This issue has been patched in version 1.6.3.
JupyterHub is software that allows one to create a multi-user server for Jupyter notebooks. Prior to version 5.4.4, an open redirect vulnerability in JupyterHub allows attackers to construct links which, when clicked, take users to the JupyterHub login page, after which they are sent to an arbitrary attacker-controlled site outside JupyterHub instead of a JupyterHub page, bypassing JupyterHub's check to prevent this. This issue has been patched in version 5.4.4.
OAuthenticator: Authentication Bypass in Auth0OAuthenticator via Unverified Email Claims
OAuthenticator is software that allows OAuth2 identity providers to be plugged in and used with JupyterHub. Prior to version 17.4.0, an authentication bypass vulnerability in oauthenticator allows an attacker with an unverified email address on an Auth0 tenant to login to JupyterHub. When email is used as the usrname_claim, this gives users control over their username and the possibility of account takeover. This issue has been patched in version 17.4.0.
Missing Authentication for Critical Function in mlflow/mlflow
In mlflow/mlflow, the FastAPI job endpoints under `/ajax-api/3.0/jobs/*` are not protected by authentication or authorization when the `basic-auth` app is enabled. This vulnerability affects the latest version of the repository. If job execution is enabled (`MLFLOW_SERVER_ENABLE_JOB_EXECUTION=true`) and any job function is allowlisted, any network client can submit, read, search, and cancel jobs without credentials, bypassing basic-auth entirely. This can lead to unauthenticated remote code execution if allowed jobs perform privileged actions such as shell execution or filesystem changes. Even if jobs are deemed safe, this still constitutes a
FastMCP is the standard framework for building MCP applications. Prior to version 3.2.0, while testing the GitHubProvider OAuth integration, which allows authentication to a FastMCP MCP server via a FastMCP OAuthProxy using GitHub OAuth, it was discovered that the FastMCP OAuthProxy does not properly validate the user's consent upon receiving the authorization code from GitHub. In combination with GitHub’s behavior of skipping the consent page for previously authorized clients, this introduces a Confused Deputy vulnerability. This issue has been patched in version 3.2.0.
FastMCP has a Command Injection vulnerability - Gemini CLI
FastMCP is the standard framework for building MCP applications. Prior to version 3.2.0, server names containing shell metacharacters (e.g., &) can cause command injection on Windows when passed to fastmcp install claude-code or fastmcp install gemini-cli. These install paths use subprocess.run() with a list argument, but on Windows the target CLIs often resolve to .cmd wrappers that are executed through cmd.exe, which interprets metacharacters in the flattened command string. This issue has been patched in version 3.2.0.
pymetasploit3 - Command Injection via Newline Injection in console.run_module_with_output()
Command injection vulnerability in console.run_module_with_output() in pymetasploit3 through version 1.0.6 allows attackers to inject newline characters into module options such as RHOSTS. This breaks the intended command structure and causes the Metasploit console to execute additional unintended commands, potentially leading to arbitrary command execution and manipulation of Metasploit sessions.
In Tornado before 6.5.5, cookie attribute injection could occur because the domain, path, and samesite arguments to .RequestHandler.set_cookie were not checked for crafted characters.
vLLM: Downmix Implementation Differences as Attack Vectors Against Audio AI Models
vLLM is an inference and serving engine for large language models (LLMs). From version 0.5.5 to before version 0.18.0, Librosa defaults to using numpy.mean for mono downmixing (to_mono), while the international standard ITU-R BS.775-4 specifies a weighted downmixing algorithm. This discrepancy results in inconsistency between audio heard by humans (e.g., through headphones/regular speakers) and audio processed by AI models (Which infra via Librosa, such as vllm, transformer). This issue has been patched in version 0.18.0.