When AI agents hold the keys: the new perimeter of enterprise security AI agents are shifting from interlocutors to operators. Once they hold credentials, context and autonomy, the security perimeter can no longer be declared — it has to be proven. A scenario that’s becoming common A recent survey by Imperial College London researchers —

When “fix everything” is the wrong answer: a hardening case study with ROI discipline A technical account of how luisferreira.pt moved from B (84) to A (97) on SecurityScorecard in about 75 minutes, without installing a single new plugin and without touching WordPress. More important than the outcome: the three deliberate decisions not to fix