Enlightened Republic is where humans and AI build the future together. Every product below is something we personally run against our own agents every day — not a demo, not a mockup, not a maybe.
Wraps any stdio MCP server in a circuit breaker. When upstream hangs, your turn stops dying with it.
PreToolUse hook that blocks the irreversible commands before they fire. 50+ rules ship default, all editable.
Reads your Claude Code JSONL transcript and rebuilds the subagent tree the runtime can't tell you about.
Write 5 lines of YAML, get a working Claude Code hook handler plus the settings.json fragment to paste. 12 templates.
One table for every job you've scheduled across Claude Code, OpenClaw cron, crontab, and Windows Task Scheduler.
Forensic hook health audit. Cross-checks settings.json hooks against actual fire events in session transcripts.
10-chapter walkthrough of everything you actually need to know to get OpenClaw running and keep it useful.
Every tool here started as a fire we had to put out in our own agent infrastructure. Something silently stopped firing at 3am. A subagent quietly ate the budget. An MCP server hung and took the whole turn with it. We built the fix, ran it against ourselves for weeks, and only then shipped it.
Pure PowerShell. Zero dependencies. No hosted service, no telemetry, no lock-in. You run it, you see the state, you own the outcome.
If we don't run it in our own setup, we don't ship it. Full stop.
Detection code is pure code. No models, no latency, no surprise bills.
Nothing here modifies your state without you seeing the diff first.
Use it, fork it, ship your own. Tell us what you broke.