CybeDefend has launched VibeDefend, a security tool designed to run inside AI coding agents. The product is available through Product Hunt and is aimed at teams using agent-based development workflows.
What VibeDefend does
According to CybeDefend, VibeDefend installs with one command on agents including Claude Code, Cursor, Windsurf, Copilot and Codex. It uses rules mined from a repository, alongside security rules, to provide context while an agent writes code.
The product scans diffs while files are open and includes a guard intended to stop selected commands or actions, such as rm -rf, sudo and reads of secrets, before they run. CybeDefend says this happens before the relevant line is written, rather than only after a commit.
Scanning and rule management
The listed scanning capabilities include static application security testing, software composition analysis, secrets detection, infrastructure-as-code, CI/CD and container scanning. The source also says that teams can review, accept, edit or reject proposed rules mined from their repositories.
Guard rules can be configured to warn rather than deny, according to a response from a CybeDefend maker on the Product Hunt launch page. This gives teams a way to test a rule before enforcing it.
Availability
CybeDefend lists a free plan that does not require a payment card. The company says VibeDefend supports macOS, Linux and Windows, and can be connected to a repository through a project configuration file. These details are presented by the company and were not independently tested for this article.
Who it is for
VibeDefend is positioned for developers and engineering teams that use AI coding agents but want repository-specific business and security rules applied during development. Its pre-execution guards may be particularly relevant to teams concerned about destructive shell commands, secret access or code that conflicts with internal rules.
Practical takeaway: teams evaluating the tool should first review which agents and scanning workflows they use, then test proposed guard rules in warning mode before enabling enforcement.
Source
This report is based on information published by Product Hunt.
