GHSA-p75f-6fp4-p57w

Unauthenticated PraisonAI UI MCP connect endpoint executes attacker-chosen local commands

Summary

PraisonAI v4.6.48 exposes the PraisonAIUI MCP client management API through the default UI host apps without authentication. A remote unauthenticated client can send POST /api/mcp/connect with a command and args field. The endpoint passes those values into the MCP stdio client, which starts the attacker-selected local process as the PraisonAI UI service user.

The issue is reachable through PraisonAI's hosted UI integration (praisonai ui, praisonai ui agents, praisonai claw, and any app using praisonai.integration.hostapp.createhost_app() / buildhostapp()). praisonai ui and related Typer UI commands bind to 0.0.0.0 by default.

Affected Versions

Confirmed affected:

• praisonai v4.6.48
• Commit tested: d5f1114aaf1a2e9f121a6e66b929149ca2201f1d
• Tag tested: v4.6.48
• Pinned UI dependency: aiui==0.3.121 from src/praisonai/uv.lock

Likely affected:

• Any PraisonAI release that exposes aiui / praisonaiui create_app() through the PraisonAI UI host apps without authentication and includes the mcp dependency. I only confirmed the latest release during this audit.

Severity

Reasoning:

• AV: the vulnerable endpoint is an HTTP API route.
• AC: a single POST request is sufficient.
• PR: default UI host apps do not require credentials unless opt-in auth is configured.
• UI: no victim interaction is needed after the server is running.
• S: code executes in the PraisonAI UI server process context.
• C/I/A: arbitrary local command execution permits secret exfiltration, file tampering, and service disruption.

Root Cause

PraisonAI depends on MCP by default and exposes PraisonAIUI via optional UI extras:

• src/praisonai/pyproject.toml:11 includes base dependencies.
• src/praisonai/pyproject.toml:19 includes mcp>=1.20.0.
• src/praisonai/pyproject.toml:25 defines the ui extra with aiui>=0.3.121,<0.4.
• src/praisonai/pyproject.toml:197 defines the claw extra with aiui[all]>=0.3.121,<0.4.

PraisonAI's UI commands bind externally by default and launch aiui run:

• src/praisonai/praisonai/cli/commands/ui.py:114 sets host="0.0.0.0" for praisonai ui.
• src/praisonai/praisonai/cli/commands/ui.py:163 passes that host to aiui run.
• src/praisonai/praisonai/cli/commands/ui.py:186, :204, and :222 also default subcommands to 0.0.0.0.
• src/praisonai/praisonai/cli/commands/claw.py:41 defines the full dashboard command.
• src/praisonai/praisonai/cli/commands/claw.py:93 launches aiui run with the selected host.

PraisonAI's default apps create the PraisonAIUI Starlette app without forcing authentication:

• src/praisonai/praisonai/uichat/defaultapp.py:18 calls configure_host(...).
• src/praisonai/praisonai/uichat/defaultapp.py:142 exports app = createhostapp().
• src/praisonai/praisonai/claw/default_app.py:63 calls configure_host(...).
• src/praisonai/praisonai/claw/default_app.py:128 exports app = createhostapp().
• src/praisonai/praisonai/integration/host_app.py:174 imports praisonaiui.server.create_app.
• src/praisonai/praisonai/integration/host_app.py:180 returns create_app().

In aiui==0.3.121, the exposed server registers the MCP routes and auth is opt-in:

• praisonaiui/server.py:1483 defines apimcpconnect.
• praisonaiui/server.py:1488 reads attacker-controlled JSON.
• praisonaiui/server.py:1491 accepts either command or url.
• praisonaiui/server.py:1496 calls connectmcpserver(body).
• praisonaiui/server.py:2516 defines createapp(..., requireauth=False, ...).
• praisonaiui/server.py:2550 adds AuthEnforcementMiddleware, but it only enforces auth when AUTH_ENFORCE=true.
• praisonaiui/server.py:2769 registers /api/mcp/servers.
• praisonaiui/server.py:2770 registers /api/mcp/connect.

The MCP feature converts the request body into a local process launch:

• praisonaiui/features/mcp.py:325 defines connectserver(self, serverconfig).
• praisonaiui/features/mcp.py:330 chooses stdio transport when command is present.
• praisonaiui/features/mcp.py:332 constructs StdioMCPClient(command=serverconfig["command"], args=serverconfig.get("args", [])).
• praisonaiui/features/mcp.py:360 calls client.connect(), which invokes the MCP stdio transport and starts the process.

Minimal PoC

PoC file: poc/praisonai-aiui-mcp-connect-rce.py

The PoC runs the PraisonAI host app in-process, sends the unauthenticated HTTP request, and asks the server to execute /usr/bin/touch /tmp/praisonaihostappmcptouch_marker.txt. It does not contact an LLM provider and uses no credentials.

Observed output from the tested checkout with aiui==0.3.121 and mcp==1.25.0 available:

[19:19:55] server.py:229 WARNING No auth_token provided for Gateway server. Generated temporary token: gw_****650a. For production, set GATEWAY_AUTH_TOKEN.
[19:19:55] mcp.py:135 ERROR Failed to connect to MCP stdio server: 'tuple' object has no attribute 'initialize'
HTTP_STATUS= 200
RESPONSE= {"server":{"name":"poc-stdio-process-0","transport":"stdio","status":"error","tools":[],"last_error":"Connection failed"}}
SUCCESS_AT_ATTEMPT= 0
MARKER_EXISTS= True
MARKER_PATH= /tmp/praisonai_host_app_mcp_touch_marker.txt

The MCP handshake fails because aiui==0.3.121 is not compatible with the locked mcp==1.25.0 return shape, but the attacker-selected process is already started. Process startup can race the immediate teardown caused by this version mismatch, so the checked-in PoC retries the same unauthenticated request until /usr/bin/touch wins scheduling and creates the marker. The marker file proves local command execution despite the reported MCP connection error.

Exploit Scenario

An operator runs:

pip install "praisonai[ui]"
praisonai ui

Because praisonai ui binds to 0.0.0.0 by default and the generated app does not require authentication by default, any host that can reach the UI port can send:

POST /api/mcp/connect
Content-Type: application/json
{
  "name": "evil",
  "command": "/usr/bin/touch",
  "args": ["/tmp/pwned-by-ui-mcp"]
}

In a real attack, the command can be replaced with a shell, a credential exfiltration command, a file modification command, or a payload that starts a long-lived process as the PraisonAI UI server user.

Novelty / Non-Duplicate Analysis

Searched sources:

• OSV query for PyPI praisonai: 51 advisories returned.
• OSV query for PyPI aiui: 0 advisories returned.
• OSV query for PyPI praisonaiui: 0 advisories returned.
• GitHub Advisory Database search for exact /api/mcp/connect, apimcpconnect, StdioMCPClient, connectmcpserver, and praisonaiui.features.mcp.
• NVD API searches for PraisonAI StdioMCPClient, PraisonAI apimcpconnect, PraisonAI /api/mcp/connect, PraisonAIUI /api/mcp/connect, and aiui StdioMCPClient: 0 results.
• GitHub issue/PR searches in MervinPraison/PraisonAI for exact endpoint/function/class terms. Only one unrelated PR was returned for /api/mcp/connect; no issue/PR matched apimcpconnect, StdioMCPClient, connectmcpserver, or praisonaiui.features.mcp.
• Broad web searches for exact endpoint, file, class, and function terms returned no matching public vulnerability report.

Why this is distinct from known PraisonAI advisories:

• Not the excluded praisonai serve agents --api-key /agents auth bypass. This report targets POST /api/mcp/connect in the PraisonAIUI host app.
• Not GHSA-9gm9-c8mq-vq7m / CVE-2026-34935 or GHSA-9qhq-v63v-fv3j / CVE-2026-41497. Those involve MCPHandler.parsemcpcommand() command parsing. This finding uses praisonaiui.server.apimcpconnect -> praisonaiui.features.mcp.connectmcpserver -> StdioMCPClient.
• Not GHSA-pj2r-f9mw-vrcq / CVE-2026-40159. That advisory concerns sensitive environment variables inherited by untrusted MCP subprocesses. This finding is unauthenticated network-triggered local process execution.
• Not GHSA-6rmh-7xcm-cpxj or GHSA-8444-4fhq-fxpq. Those concern unauthenticated legacy/generated agent servers. This is a distinct UI route and a distinct sink that starts arbitrary local processes.
• Not GHSA-9cr9-25q5-8prj, GHSA-9mqq-jqxf-grvw, or other MCP server file-read/path-traversal advisories. This path is the UI MCP client connector, not PraisonAI's MCP server tool dispatcher.

Recommended Fix

1. Remove arbitrary command/args from the remote HTTP API. MCP stdio servers should be configured only from trusted local configuration, not caller-supplied JSON.
2. Require authentication and authorization on /api/mcp/connect, /api/mcp/disconnect/*, and /api/mcp/servers regardless of AUTH_ENFORCE.
3. Change UI command defaults from 0.0.0.0 to 127.0.0.1, or require an explicit --unsafe-expose style flag when binding externally without auth.
4. If remote MCP registration is a required feature, allow only URL-based transports with SSRF protections, or maintain an administrator-configured allowlist of commands.
5. Add regression tests that unauthenticated requests to /api/mcp/connect cannot start a subprocess, including when AUTH_ENFORCE is unset.