Agent access & guardrails
The AI agents operating WealthEngine each get their own scoped identity — never the keys to the whole app — governed by least privilege, a trust dial, per-action guardrails, a full audit trail, and an instant kill switch.
Monitored systems · your live apps
Checking live status…Live read-only health of your deployed apps. Warden authorizes agent actions against them — it never executes the risky ones.
Agents
Three Bits AI agents that operate WealthEngine, across the trust gradient. Open one to set its scope, autonomy, and guardrails.
Cost SentinelActs on its own — within scope + guardrails
ActiveAutonomous5 grants3 guardrails
Owner: Platform Team
Bits AI SREActs, but every action needs approval
ActiveSupervised8 grants2 guardrails
Owner: Priya Nair
Bits AI Security AnalystSuggests only — never acts
ActiveAssist3 grants1 guardrails
Owner: Security Team