Skip to main content
HG

A local approval layer for risky AI agent actions

HOL Guard sits between AI coding agents and your machine, pausing risky actions for approval before they run. It can intercept shell commands, reads of secret-bearing files, package installs, and MCP server registrations while allowing routine work to continue.

Agenticness = how independently a tool can take action, scored across 9 dimensions. Scored independently by David Kooi, Skylark Creations — see full rubric → Last re-evaluated 10 September 2026.

See pricing ↓

Desktop
Code Execution
File Access
B2B
MCP Support
On-Device / Edge
Copilot (Human-in-Loop)
Visit HOL Guard

Is this your tool? Claim this listing to manage your content and analytics.

Ask about HOL Guard

Get answers based on HOL Guard's actual documentation

Try asking:

About

What It Is

HOL Guard is a local security and control layer for developers using AI agents that can run tools on their computer. Rather than acting as an agent itself, it monitors the actions requested by compatible coding assistants and requires your approval before potentially dangerous operations execute.

What to Know

Guard's value is in prevention before execution: it is designed to stop an agent from reading credentials, deleting files, installing packages, exfiltrating environment-file contents, or pushing directly to a protected branch. The site says checks...

Key Features
Intercepts risky AI-agent actions before they execute
Prompts you to approve or deny dangerous shell commands
Blocks or pauses reads of secret-bearing files such as .env, .npmrc, and cloud credentials
Wraps npm, pip, and Cargo installation commands and records the decision
Pauses remote MCP endpoint registrations before they are added
Use Cases
Review an AI coding agent's attempted access to AWS credentials while debugging a production deployment
Block an agent from deleting a documents or project directory with a destructive rm command
Approve a package installation only after reviewing the agent-generated npm, pip, or Cargo command
Agenticness: Guided Assistant

Executes tasks you assign, one step at a time, within narrow domains.

High evidence
Last evaluated: Sep 10, 2026

Dimension Breakdown

Action Capability
Autonomy
Planning
Adaptation
State & Memory
Reliability
Interoperability
Safety
Operator Sovereignty

Categories

Pricing
  • Free: Free to start; the site also states a no-account-required option is available.
  • Pro: Mentioned as an upgrade option with saved setup, receipt syncing across machines, curated advisories, and Slack and email alert management. Price not provided.
Details
AddedSeptember 10, 2026
RefreshedSeptember 10, 2026
Agenticness
Quick Facts
DeploymentOn-device / local
AutonomySemi-autonomous
Model supportMulti-model
MCP supportYes
Team supportIndividual only
Pricing modelFreemium
Interfacecli
Sources
Last updated September 10, 2026
Free daily brief

When HOL Guard changes its pricing, we catch it the next morning

We recrawl every tool in this directory daily. AI Insider is the 3-minute brief on what actually changed — pricing, features, score moves on our 36-point rubric — often before any press coverage. Sent daily. Skipped when nothing happened.

I use AI for: