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 186-190 of 235 articles.
Engineering
Jungle Grid gives AI agents 9 MCP tools to submit, monitor, and retrieve GPU workloads across 247 nodes in 18 countries — with automatic VRAM fit checks, cross-provider failover, and managed artifact storage. No GPU knowledge required. $3 in free credits.
Product
Most AI agents start fresh every session. Facio agents update their own memory files — MEMORY.md, BUGS.md, USER.md — in real time during conversations, using surgical edits that keep the knowledge base clean and current. Here's how inline learning works, when agents write to memory, and why this architecture produces better long-term behavior than RAG alone.
Security
Non-human identities outnumber employees 144:1 in enterprise environments. 78% of employees use unauthorized AI tools. 97% of machine identities have excessive privileges. Shadow AI agents aren't a future threat — they're inside your organization right now, operating without anyone's approval.
Human-in-the-loop
Putting HITL approval rules in code creates rigid, slow-to-change systems. Putting them in prompts creates unreliable, bypassable suggestions. The right place is declarative, version-controlled configuration — a policy manifest the agent runtime enforces deterministically.
Engineering
Ejentum exposes four cognitive harnesses (reasoning, code, anti-deception, memory) as MCP tools that inject structured scaffolds into LLM context — catching sycophancy, hallucination, and causal shortcuts before they reach users. Benchmarked at +10.1pp across 180 tasks.