Observed arrival · 2026-09-12
Stroq, the firewall between an AI agent and its next command
An open-source local security layer that scans what AI coding agents read, marks risky sessions, and blocks dangerous follow-up actions.
○Open
⊠Login
$Paid
†Ads
✦Pretty
●Pro
◎Niche
◉Human
⚑Risk
ƒJS
Field notes
The project uses an inbound scan-and-taint phase and an outbound action-judgment phase, with adapters for named coding agents and a stdio proxy for other MCP clients. The homepage reports roughly 100–250 ms per hook, 1,660 tests, and 599 gated rules, including vendored threat rules. It also describes a fail-closed response to engine errors and an audit chain intended to record blocked or evaluated actions.
Observed signals
Read the marks
Editorial observations of this landing page, not a rating.
○OpenPublic substance visible
✦PrettyNotable craft visible
●ProPolished or operationally mature
◎NicheUnusually specific use
One card from the complete issue