Guides

Guides for putting AI agents to work safely

How to verify what agents did, require the right approvals, secure MCP tool calls, keep evidence, retry safely and measure reliability.

How to verify what an AI agent actually did

A practical method to verify AI agent actions: postconditions, independent read-back checks, honest outcomes and receipts instead of trusting the API.

8 min read

Human-in-the-loop approvals for AI agents

Add human approvals to AI agent actions without loopholes: risk-based policy, payload-bound approvals, expiry, separation of duties and dual control.

7 min read

Securing MCP tool calls in production

How to secure Model Context Protocol tool calls in production: scoped keys, deny-by-default policy, approvals, egress controls and verified outcomes.

8 min read

Audit trails and tamper-evident receipts for AI agents

What an AI agent audit trail should record, how hash-chained receipts make edits detectable, and what they can and cannot prove.

7 min read

Safe retries and idempotency for AI agent actions

Why naive retries duplicate payments and records in agent systems, and how to retry safely with idempotency keys and verification of unknown outcomes.

7 min read

How to evaluate AI agent reliability

Benchmark AI agents on real outcomes: reproducible suites, fault injection, verification status, policy compliance, retries, latency and methodology.

7 min read