Dashboard

DEMO v2 · LIVE API Home control layer above your AI agents

No active project yet

You're signed in, but you don't have a workspace yet.

Active agents
Awaiting approval
Audit records
Chain integrity

Agents

Build anywhere — Zapier · n8n · Make · custom. Control here.

Anomaly detection

Heuristic — flags unusual agent behavior for human review

How to break it

Tamper-evidence, demonstrated
Editing any audit row changes its hash — verification then fails from that row onward. Tampering is detectable by anyone with the public key. No need to trust us.
  1. Click 🔒 Verify chain above → every row shows ✓, KPI turns ✓.
  2. Edit one row as an "attacker with DB access":
    # local file store: npm run tamper # cloud (Supabase SQL editor): DROP RULE IF EXISTS audit_log_no_update ON audit_log; UPDATE audit_log SET description='TAMPERED' WHERE chain_position=5 AND org_id='11111111-1111-4111-8111-111111111111'; CREATE OR REPLACE RULE audit_log_no_update AS ON UPDATE TO audit_log DO INSTEAD NOTHING;
  3. Click 🔒 Verify chain again → row 5 and everything after it turn , the integrity KPI flips to ✗, red banner shows expected vs got.
  4. Restore row 5's original description (Subscription renewal €40) → the chain verifies ✓ again.
SAFE PREVIEW Client-side only — touches no data. For when you can't run the DB step live.