About AgentGuard
AgentGuard focuses on security decisions around agent actions, components, and developer workflows. GoPlus lists AgentGuard as an AI agent security solution, and the public repository describes an open-source runtime security layer for AI coding agents.
AI agents can call tools, access files, execute commands, connect to networks, and depend on skills, plugins, packages, and MCP servers. AgentGuard is focused on making security decisions around those actions and dependencies before they become trusted parts of a workflow.
Tool actions
File access
Commands
Network requests
Agent components
What AgentGuard Publishes Today
Evaluate documented high-risk actions before execution and apply the available policy decision path.
Explore Runtime Guard
Review agent components for documented risks including prompt injection, malicious tools, credential leaks, and backdoors.
Explore Deep Scan
Check an OpenClaw workspace for suspicious skills, modified plugins, new MCP servers, and drift in trusted files.
Explore OpenClaw Patrol
From Component Review to Runtime Decisions
Scan relevant skills, plugins, MCP servers, agents, and related dependencies within the documented scope.
Use the available integration path to assess high-risk actions before execution.
Different hosts use different integration modes, and complete third-party MCP runtime coverage is not claimed.
Review Security Boundaries
npm install -g @goplus/agentguardThe public AgentGuard repository describes an open-source runtime security layer for AI coding agents and lists the MIT license. Developers can review the source, install the published package, and follow the current Docs for supported integration paths.
Read the Docs
View the Repository
[confirmed public GitHub URL]
Review the documented product and implementation paths, or contact AgentGuard with a scoped evaluation or partnership request.
Explore AgentGuard
/
Contact AgentGuard
/
official_homepage
gopluslabs_home
github_agentguard_repo
github_agentguard_readme_raw
github_agentguard_package_json
Review the security position or contact AgentGuard for information that requires current confirmation.