tag
Read mini classes on this topic together, starting with the latest lessons for context.
lessons
Tag guide
This tag gathers mini classes around one specific topic. Start with the newest post, then use categories and related reading for context.
A scan score measures how severe a flaw is, not how urgent it is for you. Ask: is it exploited, does your service run that code, can outside input reach it? AI gathers evidence; people close alerts.
An AI agent on an employee's ERP account leaves only that name in the logs, so no one can tell who changed what or stop just the AI. Give it its own identity, deny by default, and escalate above set amounts.
When an AI workflow tool is exposed, the danger is the credentials stored in each flow. Close entry points, preserve logs, rotate by which flow held which key, and reissue with less access.
Malicious AI agent extensions can hide harmful parts until runtime and slip past over 90% of install-time scans. Before installing, run one with no real data, network, or write access, and see what it touches.
Claude runs in Microsoft Foundry, but Anthropic still processes prompts, flagged content may get human review, and some setups can leave your region. Ask: who processes data, where, and who sees it?
PamStealer, Mac malware posing as a clipboard tool, asks for your password and keeps asking until it's right. Looks can't tell real from fake; context can. If you can't explain why it's asking, press Cancel.
AI now finds vulnerabilities, writes patches, and runs tests. Which patches ship without a person depends on rollback: reversible, tested upgrades can move fast; login, payment, data, or deployment changes need a human.
One click on a Microsoft link could make Microsoft 365 Copilot leak mailbox codes via an image URL. Office AI inherits every read permission a user has, so decide what stays out of its search.
Meta's AI support was talked into changing strangers' emails, and over 20,000 Instagram accounts were stolen. The fix: before AI handles accounts, identity and permissions must come from the system, not the conversation.
AI makes fake sites and texts fast and convincing; even Taiwan's official government short code has been spoofed. Instead of judging whether a text looks real, ask: can I do what it wants without tapping this link?
Leaks need three things at once: AI sees secrets, reads untrusted content, and can send data out. Lockdown Mode removes the third; hidden instructions can still skew answers. Don't combine all three.
