Facio Blog
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 1-5 of 81 articles.
Engineering
Featured article
AI Design Blueprint exposes 13 read-only MCP tools covering 10 production-tested principles for safe, observable, and steerable agent UX — with 96+ implementation examples and a paid Blueprint Readiness Score for repo-level governance review.
Product
AI agents with training data cutoffs are answering questions about a world that has moved on. Facio's web_search and web_fetch tools give agents direct access to the live web — search for sources, fetch the actual content, extract what's relevant, and reason over it. Combined with the search-reason loop, agents can run production-grade research workflows that match what humans do with browser tabs.
Security
Meta's hallucinated permission exposure, the Mercor LiteLLM supply chain breach, AI-generated polymorphic malware, coordinated AI-DDoS-API campaigns, AI agent control failures, and the model leak fallout. Each incident dissected with full MITRE ATT&CK mapping.
Human-in-the-loop
A customer-service AI agent just approved a $50,000 refund — to a fraudulent account. The CFO wants answers. Here is the four-tier HITL architecture that lets customer-facing agents move at conversational speed for routine cases and force a synchronous human checkpoint for everything that can hurt the business.
Engineering
ThinAir Data gives AI agents 23 read-only MCP tools across PostgreSQL, MySQL, and SQL Server — with three independent safety layers (SQL guard, session READ ONLY, per-connection firewall) and runtime connection management.