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 36-40 of 235 articles.

Product

Facio's Normalization Layer: How AI Agents Handle the Messy Real-World Inputs That Never Match the Schema

Jul 15, 2026Product

Facio's Normalization Layer: How AI Agents Handle the Messy Real-World Inputs That Never Match the Schema

Production AI agents receive inputs from dozens of sources: web forms, APIs, webhooks, email, chat, file uploads, copy-paste from spreadsheets. Every source has its own format, conventions, and quirks. Every source produces messy data: missing fields, wrong types, locale variations, encoding issues, ambiguous values. Naive agents assume clean inputs and fail on real-world data. Facio's normalization layer sits between the world and the agent's reasoning — cleaning, standardizing, and validating every input before the agent sees it. The agent reasons over clean data; the team reasons about fewer failures.

Security

SPIFFE Workload Identity for AI Agents: The Cryptographic Foundation That Makes Zero Trust Possible

Jul 15, 2026Security

SPIFFE Workload Identity for AI Agents: The Cryptographic Foundation That Makes Zero Trust Possible

Static API keys, shared secrets, and long-lived tokens cannot serve AI agents at machine velocity. SPIFFE's X.509 and JWT SVIDs — cryptographically verifiable, automatically rotated, scoped to the workload — are the cryptographic foundation for zero trust AI agent deployment. Vault integration went GA in June 2026.

Human-in-the-loop

HITL as a Forgetting Curve: Why Reviewer Memory Is the Limiting Factor Nobody Measures

Jul 15, 2026Human-in-the-loop

HITL as a Forgetting Curve: Why Reviewer Memory Is the Limiting Factor Nobody Measures

Reviewers forget. The first 50 actions of the day are calibrated. The next 100 are fatigued. The last 50 are rubber stamps. HITL systems are designed for the first 50, monitored for the middle 100, and silent about the last 50. The forgetting curve is the limiting factor on HITL quality — and most teams don't measure it because measuring it would force them to redesign the schedule.

Engineering

MCP Spotlight: Filesystem MCP Server — Anthropic's Reference Implementation for Sandboxed, Allowed-Directory-Scoped File Access

Jul 14, 2026Engineering

MCP Spotlight: Filesystem MCP Server — Anthropic's Reference Implementation for Sandboxed, Allowed-Directory-Scoped File Access

The official Filesystem MCP Server by Anthropic — ~14 focused tools covering read, write, edit, search, directory, metadata. Explicit allowed-directories allow-list at startup (deny-by-default), Roots-aware dynamic updates, image support. MIT-licensed, available via NPX or Docker. The local-file-access default for AI agents in 2026.

Product

Facio's Token Discipline: How AI Agents Spend Their Context Budget Like Engineers Spend Cloud Credits

Jul 14, 2026Product

Facio's Token Discipline: How AI Agents Spend Their Context Budget Like Engineers Spend Cloud Credits

An AI agent's context window is its working memory — and like cloud credits, it has a budget. A team that ignores the budget wakes up to a $50k monthly bill; an agent that ignores the context budget forgets, hallucinates, and produces degraded output by mid-session. The naive approach is to give the agent whatever context it asks for and hope for the best. Facio's token discipline treats the context window as a managed resource: tracked, allocated, optimized, and accounted for. The agent stays sharp; the team stays within budget.