CVE-2026-102610: fast-agent: `fast-agent serve` binds HTTP to 0.0.0.0 with no authentication by default; with --shell this is unauthenticated remote command execution

Published Nov 6, 2026
·
Updated

fast-agent is a framework for building, running, and evaluating AI agents with Model Context Protocol support. Prior to 0.7.21, fast-agent serve uses HTTP transport bound to 0.0.0.0 on port 8000 without authentication, allowing any network client that can reach the /mcp endpoint to invoke the exposed agent tool. When the server is started with --shell, the agent can call the local execute tool implemented by tools/localshellexecutor.py, so an unauthenticated remote client can cause operating-system commands to run as the server user; without --shell, the client can still consume the operator's model and configured MCP tools. The defaults are defined in cli/commands/serve.py and mcp/server/agentserver.py, and the non-interactive server permission prompt does not prevent the described access. This issue is fixed in version 0.7.21.

Remediation

Recommended actions to resolve this vulnerability, in priority order.

  1. Upgrade

    Upgrade fast-agent to a version that resolves this vulnerability.

    Fixed in 0.7.21

Event History

Nov 6, 2026
CVE Published
via MITRE·08:55 PM
Data Sourced
via MITRE·08:55 PM
DescriptionSeverityWeakness
Data Sourced
via NVD·09:17 PM
DescriptionSeverityWeakness

Frequently Asked Questions

1

Which deployments are exposed to remote command execution?

Deployments running a version before 0.7.21 are exposed if they use fast-agent serve with --shell and an attacker can reach the /mcp endpoint over the network. The service binds to 0.0.0.0 on port 8000 without authentication by default, allowing commands to run as the server user.

2

Is a server still at risk if it was not started with --shell?

Yes. Without --shell, an unauthenticated reachable client can still invoke the exposed agent tool and consume the operator's model and configured MCP tools. The described local operating-system command execution path specifically requires --shell.

3

Does the non-interactive server permission prompt block unauthenticated access?

No. The non-interactive server permission prompt does not prevent the described access to the /mcp endpoint.

4

How can I determine whether my instance is affected?

Check whether the deployment runs a fast-agent version earlier than 0.7.21 and whether it uses fast-agent serve. Also determine whether port 8000 and the /mcp endpoint are reachable by untrusted network clients, and whether the service was started with --shell.

5

What is the available fix?

Upgrade fast-agent to version 0.7.21, which fixes this issue.

Contact

SecAlerts Pty Ltd.
132 Wickham Terrace
Fortitude Valley,
QLD 4006, Australia
info@secalerts.co
By using SecAlerts services, you agree to our services end-user license agreement. This website is safeguarded by reCAPTCHA and governed by the Google Privacy Policy and Terms of Service. All names, logos, and brands of products are owned by their respective owners, and any usage of these names, logos, and brands for identification purposes only does not imply endorsement. If you possess any content that requires removal, please get in touch with us.
© 2026 SecAlerts Pty Ltd.
ABN: 70 645 966 203, ACN: 645 966 203