Blog
AI security & governance, for humans
Practical writing on Shadow AI, runtime controls, evidence, and red-team validation — and how Argorix turns those problems into operating workflows.
Shadow AI
Shadow AI discovery: why repository scans are not enough
Modern AI usage spreads across repositories, browser sessions, pipelines, prompts, models, and external tools. To govern it, you need visibility across the full estate — not just code.
Runtime Controls
Runtime guardrails for classic and agentic AI
How prevention, runtime decisioning, and telemetry differ when teams move from fixed AI apps to autonomous workflows.
Evidence Hub
Reusable evidence is the missing layer in AI compliance
Validation, findings, and reports should produce audit-ready evidence instead of forcing teams to rebuild proof.
Red Team
A practical taxonomy of AI red-team attacks
Prompt injection, instruction leakage, data exposure, model manipulation — and how to validate each against your systems.
Shadow AI
Building an AI inventory your auditors will actually trust
Owners, models, prompts, datasets, providers — the fields that turn a list of tools into a governable inventory.
Runtime Controls
Prompt injection is not going away. Plan for it.
Why input filtering alone fails, and how layered runtime controls reduce blast radius for injection attacks.
Governance
The executive AI-risk report that gets read
What boards actually want from an AI governance update — and how to produce it from live evidence.