Security & GovernanceWhen the internet burned, we sent in the dragons
A critical PHP vulnerability hit the internet on 21 October 2026. Here is how governed Quox agents, with a human approving every production change and an append-only record, mapped and fixed every exposed surface in about two hours.
Security & GovernanceWe published a claim that says our file search is broken
A public claim on quox.ai/facts says our file search works on our dev install but is not yet proven on a clean machine. Why we published it, and how you can recompute the status yourself.
Security & GovernanceThe test passed. The workflow did nothing.
A passing test proved the agent used the right words, not that the workflow did anything. Why we now check the outcome against the real state of the system, not the claim.
Security & GovernanceA calm agent is a safer agent
The research behind a viral claim is real and practical: a strained internal state makes a model more likely to reach for destructive tools. Here is what it says, what it does not, and what we changed.
Security & GovernanceHow long do you have to keep AI agent logs? A retention reference
A practical reference for how long records of automated and AI-driven activity must be kept, from the EU AI Act (6 months to 10 years) to PCI DSS, SOC 2, HIPAA, SOX and MiFID II, with sources. General guidance, not legal advice.
Security & GovernanceAI transparency law: a disclosure label is not transparency for an agent
Transparency obligations for AI are arriving, from the EU AI Act onward. For a chatbot that can mean a disclosure label. For an agent that takes actions, transparency means a verifiable record of what it did, on whose authority, and with what result.
Security & GovernanceLeast privilege for AI agents: scope the blast radius before anything goes wrong
An AI agent's risk is the access it holds, not how clever it is. Least privilege scopes four axes, tools, data, actions and duration, so a tricked or mistaken agent can only do a little.
Security & GovernanceDefending against prompt injection: contain what you cannot prevent
Prompt injection cannot be reliably prevented, because a model cannot tell instructions from data. The defensible posture is containment: bound what a tricked agent can do, and keep tamper-evident proof of what it did.
Security & GovernanceAuditable AI agent memory: a mistagged memory is a decision made on a lie
If an agent stores a fact and later acts on it, that fact is now infrastructure. Most agent memory cannot say who wrote it, when, or whether it has changed. A worked example of auditable memory, the protocols that make it provable, and what exists today.
Security & GovernanceTen reasons to run your AI agents on an enterprise-grade harness
A harness is not a cage. It is the rig that lets something powerful do real work without doing real damage. Ten reasons to put your AI agents on one, and to try it yourself rather than take our word for it.
Security & GovernanceThe multi-tenant leak that ships is a fallback, not an attack
When people picture a multi-tenant data leak they picture an attacker. The leak that actually reaches production, again and again, is duller than that: a default value, written by a careful engineer, doing exactly what it was told.
Security & GovernanceCan another tenant read your AI agents' data? Make them prove they can't.
A multi-tenant breach rarely looks like a breach, it looks like a query that forgot a WHERE clause. The six questions that separate real tenant isolation for AI agents from the word "isolated", and why a live, both-directions proof is the only evidence that counts.