🤖 AI agents lab
An AI agent is a non-human identity (NHI) — its own row, its own scopes, its own secret, and one central switch that kills its access everywhere the instant you flip it.
Every agent's secret is stored only as a SHA-256 hash — the plaintext is shown to you ONCE at
registration. Authenticating a secret always goes through the SAME function
(authenticateAgent), so later agent features (delegation, tool
calls, agent-to-agent chains) built on this registry inherit the kill switch for free.
1. Register an agent
Give it a name and the scopes it should be allowed to act with.
2. Your agents
3. Test agent access
Paste an agent's id and secret to run it through authenticateAgent —
the same call every later agent feature makes. Register → test (active) → kill → test again (dead).