Monitor vulnerabilities like this one.
Sign up free to get alerted when software you use is affected.
10.0
CVE-2026-53710: ContextForge python sandbox can execute arbitrary code
CVE-2026-53710 · published 10 days ago
Summary
The python sandbox used by ContextForge’s MCP server can be tricked into running any command because it exposes a function that bypasses its safety checks. An attacker could send specially crafted code over the network and cause the server to run system commands. Update the software to a version that removes the unsafe function or add proper authentication and input validation.
What to do
- Update mcp-contextforge-gateway to version 1.0.2.
Affected software
| Ecosystem | Vendor | Product | Affected versions |
|---|---|---|---|
| pip | – | mcp-contextforge-gateway |
<= 1.0.1 Fix: upgrade to 1.0.2
|
| PyPI | – | mcp-contextforge-gateway |
< 1.0.2 Fix: upgrade to 1.0.2
|
Original advisory text
mcp-contextforge-gateway has RestrictedPython sandbox bypass via getattr builtin in python_sandbox_server
**Commit:** `f855e54d5b7bc1c91b977574a03b91eff6b86bb6`
**Component:** `mcp-servers/python/python_sandbox_server/src/python_sandbox_server/server_fastmcp.py`
## Vulnerability
RestrictedPython's sandbox in ContextForge's `python_sandbox_server` sub-project allows arbitrary Python execution via three compounding weaknesses:
1. Raw `getattr` is exposed in `safe_builtins`, bypassing `_getattr_` mediation.
2. `validate_code` checks for literal dangerous dunder strings, but the payload constructs those names at runtime.
3. The `execute_code` MCP tool can be exposed over HTTP/SSE transport with no authentication layer.
## Proof of Concept
The local PoC was executed against the real pinned sandbox code path. It constructs dunder names at runtime, walks Python's class hierarchy through the exposed `getattr`, finds `subprocess.Popen`, and executes a harmless marker command.
Transcript excerpt:
```text
POC: IBM/mcp-context-forge RestrictedPython getattr sandbox bypass
validation={'valid': True, 'message': 'Code passed validation', 'warnings': None}
success=True
stdout="FOUND_POPEN\nb'IBM_SANDBOX_POC_PASS'\n"
IBM_RESTRICTEDPYTHON_GETATTR_POPEN_REPRO_PASS
```
PoC artifact hashes:
```text
beb2d856c5a1b7c10005c9c0fccaf3d491d4fbcc2a69bc54b5e95c5730c0874b run.py
1037dcb4f0831a3c96ca728527f939af1d3e19514b23ef54f663707875748ff8 run.sh
7026e676ee31916886f10feed959ffdbe486fda05b17421387e4717acbf90ba6 transcript.txt
```
## Production Deployment Context
ContextForge's documentation describes registering MCP servers as gateways in a running ContextForge instance. If this sandbox server is reachable over HTTP in that pattern, any client that can reach the endpoint can submit the payload through `tools/call`.
## Impact
An attacker with access to the `python_sandbox_server` HTTP endpoint can execute OS commands with the privileges of the server process. In containerized deployments with host mounts or internal network reachability, this can become filesystem compromise or network pivoting from the gateway host.
## Suggested Fix
Remove raw `getattr` and `setattr` from `safe_builtins`, expose only policy-controlled attribute access if dynamic lookup is required, replace substring filtering with a stricter RestrictedPython policy pipeline, and add authentication to HTTP transport for all `tools/call` requests.
## Caveats
This report covers the `python_sandbox_server` sub-project. The core ContextForge gateway and proxy components are not directly affected. The Critical score assumes HTTP reachability; stdio-only deployments reduce the attack vector.
**Component:** `mcp-servers/python/python_sandbox_server/src/python_sandbox_server/server_fastmcp.py`
## Vulnerability
RestrictedPython's sandbox in ContextForge's `python_sandbox_server` sub-project allows arbitrary Python execution via three compounding weaknesses:
1. Raw `getattr` is exposed in `safe_builtins`, bypassing `_getattr_` mediation.
2. `validate_code` checks for literal dangerous dunder strings, but the payload constructs those names at runtime.
3. The `execute_code` MCP tool can be exposed over HTTP/SSE transport with no authentication layer.
## Proof of Concept
The local PoC was executed against the real pinned sandbox code path. It constructs dunder names at runtime, walks Python's class hierarchy through the exposed `getattr`, finds `subprocess.Popen`, and executes a harmless marker command.
Transcript excerpt:
```text
POC: IBM/mcp-context-forge RestrictedPython getattr sandbox bypass
validation={'valid': True, 'message': 'Code passed validation', 'warnings': None}
success=True
stdout="FOUND_POPEN\nb'IBM_SANDBOX_POC_PASS'\n"
IBM_RESTRICTEDPYTHON_GETATTR_POPEN_REPRO_PASS
```
PoC artifact hashes:
```text
beb2d856c5a1b7c10005c9c0fccaf3d491d4fbcc2a69bc54b5e95c5730c0874b run.py
1037dcb4f0831a3c96ca728527f939af1d3e19514b23ef54f663707875748ff8 run.sh
7026e676ee31916886f10feed959ffdbe486fda05b17421387e4717acbf90ba6 transcript.txt
```
## Production Deployment Context
ContextForge's documentation describes registering MCP servers as gateways in a running ContextForge instance. If this sandbox server is reachable over HTTP in that pattern, any client that can reach the endpoint can submit the payload through `tools/call`.
## Impact
An attacker with access to the `python_sandbox_server` HTTP endpoint can execute OS commands with the privileges of the server process. In containerized deployments with host mounts or internal network reachability, this can become filesystem compromise or network pivoting from the gateway host.
## Suggested Fix
Remove raw `getattr` and `setattr` from `safe_builtins`, expose only policy-controlled attribute access if dynamic lookup is required, replace substring filtering with a stricter RestrictedPython policy pipeline, and add authentication to HTTP transport for all `tools/call` requests.
## Caveats
This report covers the `python_sandbox_server` sub-project. The core ContextForge gateway and proxy components are not directly affected. The Critical score assumes HTTP reachability; stdio-only deployments reduce the attack vector.
References
- https://github.com/IBM/mcp-context-forge/security/advisories/GHSA-xm98-3vcf-fph7 URL
- https://github.com/IBM/mcp-context-forge/commit/63a2900e6301b9c8a483a38d3737a1be... URL
- https://github.com/IBM/mcp-context-forge Product
- https://github.com/IBM/mcp-context-forge/releases#release-v1.0.2 URL
- https://github.com/advisories/GHSA-xm98-3vcf-fph7
Severity
10.0
Critical
Type
CWE-94Code Injection
CWE-693Protection Mechanism Failure
Timeline
Published24 Aug 2026
Updated2 Sep 2026
First seen24 Aug 2026
Monitor software like this
Free during beta