HOL Guard

Good for: Putting a local checkpoint in front of AI coding agents such as Claude Code, Codex, Cursor and OpenCode, so risky actions like deleting files, reading secrets, installing packages or sending data out are allowed, paused for your approval or blocked, with a local record of what happened

Found viaMakerhol.org/guard

Details

MakerHOL (Hashgraph Online)

Use case💻 Developer tools

Free tieryes

Starting priceFree: the open-source local runtime (Apache-2.0), no account or card needed. Optional Guard Cloud plans add sync and shared policy: Solo $4.99/mo, Pro $15/mo, Team $30/seat/mo; Enterprise on request

Pricing checkedOct 30, 2026

StatusNot tested yet

Plan we testedNot tested yet

Note: Reader submission via /submit/ (2026-10-30, maker, sent from a hashgraphonline.com address). Guard page, pricing page, GitHub repo hashgraph-online/hol-guard (Apache-2.0, about 1,000 stars, updated Oct 30, 2026) and PyPI package hol-guard 3.58.0 checked Oct 30, 2026. Installs locally (pipx or an install script) and hooks into supported agents; how much it can block depends on the agent and event type, per its own coverage pages. Download, star and transaction counts on the site are the maker's; we have not checked them. Suggested test from the submitter: ask an agent to read a .env file or run a risky shell command and check that Guard pauses it for approval.

Prices are from the official pricing page on Oct 30, 2026 and can change; check the pricing page before you buy. We haven’t hands-on tested this tool yet, so there is no score.

Links

↗Official sitehol.org/guard 🏷️Pricing pagehol.org/guard/pricing