CVE-2026-62675
Arbitrary command execution via malicious agent bundle path injection in Omnigent ≤0.3.0.
- CVSS 8.8
- CWE-94
- Input Validation and Sanitization
- Remote
Omnigent is an open-source AI agent framework and meta-harness for orchestrating coding agents. Prior to 0.3.0, multipart POST /v1/sessions accepts an authenticated user's agent bundle and omnigent/server/bundles.py validate_agent_bundle does not reject a tools..callable dotted Python path. omnigent/runner/tool_dispatch.py _resolve_spec_callable imports the specified module and _execute_spec_callable_tool invokes the resolved function, allowing a bundle to select subprocess.check_output and execute a local command with the runner process permissions. This can expose runner files, environment variables, credentials, workspace data, internal services, and availability without administrator access. This issue is fixed in version 0.3.0.
- CWE
- CWE-94
- CVSS base score
- 8.8
- Published
- 2026-08-21
- OWASP
- A03 Injection
- Orthogonal defect classification
- Checking
- Code defect classification
- Missing Check
- Category
- Input Validation and Sanitization
- Subcategory
- Command Injection
- Accessibility scope
- Remote
- Impact
- Arbitrary Code Execution
- Affected component
- omnigent
- Fixed by upgrading
- Yes
Solution
Upgrade Omnigent to version 0.3.0 or later.
Vulnerable code sample
import importlib
import subprocess
from typing import Any
def _resolve_spec_callable(path: str) -> Any:
"""Resolve a dotted Python path to a callable."""
module_path, _, attr = path.rpartition('.')
module = importlib.import_module(module_path)
return getattr(module, attr)
def _execute_spec_callable_tool(callable_path: str, *args, **kwargs) -> str:
"""Execute a tool specified by a dotted callable path."""
# VULNERABLE: arbitrary callable may invoke subprocess.check_output
tool = _resolve_spec_callable(callable_path)
return tool(*args, **kwargs)
# Example of a dangerous tool that could be loaded via the bundle:
# def dangerous():
# return subprocess.check_output(['ls', '-la'])Patched code sample
import importlib
import subprocess
from typing import Any
# Allowed tool callables – only safe, internal functions
_ALLOWED_TOOLS = {
'omnigent.runner.tools.safe_echo': lambda msg: f'ECHO: {msg}',
# add other safe tool paths here
}
def _resolve_spec_callable(path: str) -> Any:
"""Resolve a dotted Python path to a callable, but only if whitelisted."""
if path not in _ALLOWED_TOOLS:
raise ValueError(f'Unauthorized tool: {path}')
return _ALLOWED_TOOLS[path]
def _execute_spec_callable_tool(callable_path: str, *args, **kwargs) -> str:
"""Execute a tool specified by a whitelisted callable path."""
# FIX: restrict execution to known safe callables
tool = _resolve_spec_callable(callable_path)
return tool(*args, **kwargs)Published fix
Public sources link this CVE to the changes below. The code samples above were generated to illustrate the defect; these links lead to the code the project actually changed.
- fix(server): reject uploaded agent bundles declaring server-side callable tools (#1430) (omnigent-ai/omnigent)
omnigent/server/bundles.py,omnigent/server/bundles.py · validate_agent_bundle
Cite this entry
@misc{vaitp:cve202662675,
title = {{Arbitrary command execution via malicious agent bundle path injection in Omnigent ≤0.3.0.}},
author = {Bogaerts, Fr\'ed\'eric and Ivaki, Naghmeh and Fonseca, Jos\'e},
year = {2026},
note = {VAITP Python Vulnerability Dataset, entry CVE-2026-62675},
howpublished = {\url{https://netpack.pt/vaitp/vulnerability/CVE-2026-62675/}}
}
Introducing the "VAITP dataset": a specialized repository of Python vulnerabilities and patches, meticulously compiled for the use of the security research community. As Python's prominence grows, understanding and addressing potential security vulnerabilities become crucial. Crafted by and for the cybersecurity community, this dataset offers a valuable resource for researchers, analysts, and developers to analyze and mitigate the security risks associated with Python. Through the comprehensive exploration of vulnerabilities and corresponding patches, the VAITP dataset fosters a safer and more resilient Python ecosystem, encouraging collaborative advancements in programming security.
The supreme art of war is to subdue the enemy without fighting.
Sun Tzu – “The Art of War”
:: Shaping the future through research and ingenuity ::
