Facio Blog

Practical notes on human-reviewed AI agents.

Payload-powered product notes, security writing, HITL patterns, and operational guidance from the Facio runtime: long sessions, Placet approvals, audit trails, memory, providers, channels, tools, and Docker-first operations.

Showing 41-45 of 235 articles.

Human-in-the-loop

HITL as a Trust Calibration: Why the Real Job of the Reviewer Is to Calibrate the System, Not the Individual Action

Jul 14, 2026Human-in-the-loop

HITL as a Trust Calibration: Why the Real Job of the Reviewer Is to Calibrate the System, Not the Individual Action

Most HITL designs treat the reviewer's job as evaluating individual actions. But the deeper job is calibrating the system — every decision is a data point that updates the team's belief about the agent, the policy, the interface, and the reviewer pool. The reviewer who only evaluates individual actions is doing one job. The reviewer who calibrates the system is doing the job HITL was actually designed for.

Engineering

MCP Spotlight: Stripe MCP Server — The Payments Bridge With Restricted API Keys, Customer-Portal Scope Isolation, and the Money-Movement Default for Agents

Jul 13, 2026Engineering

MCP Spotlight: Stripe MCP Server — The Payments Bridge With Restricted API Keys, Customer-Portal Scope Isolation, and the Money-Movement Default for Agents

The official Stripe MCP Server by Stripe — ~30 focused tools covering customers, payment intents, subscriptions, products, invoices, refunds, disputes, payouts. Restricted API Key support for per-resource + per-action scope isolation. PCI-scope minimization, webhook-first event model. The payment-automation default for AI agents in 2026.

Product

Facio's Versioned Prompts: How AI Agent Behavior Stays Reproducible Through Model and Code Changes

Jul 13, 2026Product

Facio's Versioned Prompts: How AI Agent Behavior Stays Reproducible Through Model and Code Changes

An AI agent's behavior comes from the prompt, the model, and the code that interprets both. Change any of them and the behavior changes. The team upgrades the model — the agent's responses shift subtly. The team tweaks the prompt to fix a bug — the fix breaks three other things. The team refactors the agent code — the agent stops working entirely. Reproducibility collapses. Facio's versioned prompts discipline makes agent behavior reproducible: every prompt version is pinned, every change is tracked, every comparison is traceable.

Security

You Cannot Patch What You Cannot Red Team: Why Adversarial Testing Is Now a Production-Day Activity for AI Agents

Jul 13, 2026Security

You Cannot Patch What You Cannot Red Team: Why Adversarial Testing Is Now a Production-Day Activity for AI Agents

NIST's March 2026 red-teaming competition ran 250,000 attack attempts in five days against four frontier agents. A test that passes on Monday may fail on Friday. Traditional point-in-time penetration testing is dead for AI agents; continuous adversarial evaluation is the operating model.

Human-in-the-loop

HITL and the Cost of Asking: Why "I Need More Context" Is the Fourth Decision That Makes HITL Work

Jul 13, 2026Human-in-the-loop

HITL and the Cost of Asking: Why "I Need More Context" Is the Fourth Decision That Makes HITL Work

Most HITL designs treat "I need more context" as a request for clarification. But asking for more context is a first-class decision with its own costs, its own value, and its own failure modes. A reviewer who asks for more context is doing the work of preventing downstream errors — and the system that doesn't support the question is the system that produces uninformed decisions.