rpc.run

Diagnostic tool 9 · static control assessment

Agent-Safe RPC Readiness Checker

Assess whether an AI-agent workflow is bounded to safe reads or has the simulation, approval, secret isolation, limits, audit, and fail-closed controls needed before preparing value-moving transactions.

Select the controls that are actually enforced, then assess.

Minimum safe pattern

Start read-only. For transaction preparation, use scoped secrets, allowlists, transaction simulation, decoded human approval, bounded retries, audit records, and fail-closed execution.

Evidence boundary

This is a point-in-time self-assessment, not a live control test, secret scan, transaction simulation, or guarantee of agent safety.