Pre-execution blocking
Guard intercepts risky actions before an AI coding agent executes them, including shell commands, secret reads, MCP registrations, config changes, installs, and risky hooks.
HOL Guard
ReclamarHOL Guard is a local security layer for AI coding agents that blocks risky actions before they run, with local approval for Codex, Claude Code, Cursor, and Gemini CLI.
Local approval and scanning for AI agent actions
HOL Guard is a local security layer for AI coding agents. It is designed to stop risky actions before they execute, including shell commands, secret reads, MCP changes, and config modifications.
The product is positioned for people using AI coding tools such as Codex, Claude Code, Cursor, Gemini CLI, OpenCode, Hermes, and OpenClaw. It runs on the user’s machine, supports offline scanning, and can be used without creating an account for a single machine.
The site also distinguishes Guard from post-hoc security tools: instead of reporting risks after disclosure or after installation, it checks the action at the point an agent is about to run it. That makes it useful for both individual developers and teams that want a local approval step in front of agent-driven workflows.
What HOL Guard does
Local scanning
The product runs locally on your machine and can scan before the harness starts, which lets it work without a network call in the default setup.
Agent launch wrapping
Each supported agent gets a tailored integration that wraps the launch command, so Guard can sit in front of the agent rather than only reporting after the fact.
Inline approvals
If Guard blocks something you intended, you can review the diff on screen, allow it with one keystroke, and keep moving. The decision is recorded locally.
Receipts and alerts
When you opt in to Guard Cloud, decision receipts can be synced across machines and alerts can be managed through Slack and email.
Maintainer scanning
For maintainers, HOL also offers plugin-scanner to validate extensions in CI, with checks for MCP posture, installability, trust scoring, and marketplace metadata.
Common ways teams use HOL Guard
Developer workstation protection
Use Guard as a local gate in front of an AI coding agent so shell commands and config edits are reviewed before they run.
Fast human approval
Use the on-screen diff and one-keystroke allow flow when an AI agent attempts something legitimate but sensitive, such as a secret read or MCP registration.
Offline or disconnected workflows
Use the product on a laptop or in an environment without reliable connectivity, since local scanning works offline and does not depend on a cloud call.
Team oversight and shared receipts
Use Guard Cloud when a team wants decision receipts synced across machines and alerting through Slack or email, while keeping file contents and prompts local by default.
Extension and plugin release checks
Use plugin-scanner in maintainer CI to validate plugins before release, then use Guard locally to enforce runtime decisions after deployment.
Pros and Cons
Pros
- Runs locally and does not require a network call before scanning in the default setup.
- Works without an account on one machine and does not require a credit card for the free use case described on the site.
- Blocks risky actions before execution, including shell commands, secret reads, MCP registrations, and config changes.
- Supports several popular coding agents, with a tailored integration approach for each one.
- Provides a clear approval path when something is blocked, including on-screen diffs and one-keystroke allow actions.
Cons
- Pricing details for paid tiers are not visible in the provided source, so team and enterprise limits are unclear.
- The public page gives only partial integration and platform documentation beyond the agent list and launch-command workflow.
FAQ
Does HOL Guard work without an account?
Yes. The source says Guard is free forever on one machine with full local scanning, and it does not require sign-up, a credit card, or a cloud dependency for local protection.
Does Guard add latency?
Guard is designed to run locally and adds less than 50 ms for a typical scan, according to the site. It scans before the harness starts rather than making a network call first.
What data leaves my machine?
By default, nothing leaves your machine. The site says Guard never reads file contents server-side. If you opt in to Guard Cloud, only decision receipts are synced, not file contents, prompts, or environment variables.
Which agents work today?
Guard supports Hermes, OpenClaw, Codex, Claude Code, Cursor, Gemini CLI, and OpenCode. The site says each integration wraps the launch command so Guard can scan before execution.
Does it work offline?
Yes. Local scanning works offline. Guard Cloud syncs when you are connected and keeps the queue until it can reach the server.
Quick Facts
- Category
- Developer Tool
- Primary job
- Approve or block risky AI agent actions before execution
- Platform
- Local machine; offline capable
- Supported agents
- Hermes, OpenClaw, Codex, Claude Code, Cursor, Gemini CLI, OpenCode
- Source domain
- hol.org
- Pricing signal
- Free forever on one machine; Guard Cloud and Pro features are referenced, but plan details are not shown
Alternativas a HOL Guard
Orca
Orca is an Agent Development Environment for shipping with coding agents, running multiple CLI agents in parallel across isolated worktrees, with desktop and mobile workflows.
Firebase Studio
Firebase Studio is a web-based workspace for full-stack app development with Gemini-assisted coding, app previews, cloud emulators, collaboration, and browser deployment.
Kastra
Kastra authorization infrastructure for AI systems checks prompts, tool calls, shell commands, API requests, and browser actions before execution.
JetBrains
JetBrains makes developer tools for individuals and teams, including IntelliJ IDEA, PyCharm, DataGrip, WebStorm, Rider, and CLion, with built-in run/debug/test workflows.
小艺
小艺 is Huawei’s AI smart assistant for Q&A, writing, document reading, code help, image recognition, and file drag-and-drop.
EZsite AI
EZsite AI is an AI website builder that turns a URL into a fullstack React or Vue.js app with hosting, custom domains, code export, and backend features.