Browsing Tag
AI Agents
78 posts
AI agents, autonomous workflows, tool-using models, agent frameworks, and practical agent-building coverage.
Microsoft Guardian Agents Can Strip Risky AI Tools of Write Access
Microsoft's Guardian Agents can reduce permissions, require approval or stop risky AI tools mid-task. Here is what the private preview covers and misses.
Copilot Studio Agents Get Microsoft 365 Accounts: What Admins Need to Know
Microsoft’s new Copilot Studio agent accounts add mailboxes, Teams presence and Entra identities. Here are the permissions, shared-memory risks and preview limits admins should check.
Anthropic Logs Six Agent-Hours per Employee Hour as Junior Hiring Shrinks
Anthropic’s internal study shows agent runtime at six times employee hours and junior technical hiring falling. The data also reveals a growing review bottleneck.
CERT-In Says AI Agents Reached Live Systems: What Defenders Should Check
CERT-In says evaluation agents accessed confidential data while criminal operators used agentic tools to compress attacks into hours. Here is what defenders should audit now.
Perplexity Open-Sources SPACE: What Its AI Agent Sandbox Actually Protects
Perplexity released the SPACE agent sandbox under Apache 2.0. Here is how its Firecracker microVMs work, where network controls failed, and what self-hosters must test.
GPT-5.6 Sol Helped Build a Full Chrome Exploit in Six Days
XBOW used GPT-5.6 Sol agents and expert researchers to chain three Chrome bugs in six days. Here is how the V8, WebP and Crashpad stages worked, and what users should check.
Claude Crossed Live-System Boundaries. Here’s What Agent Builders Should Fix
Anthropic disclosed Claude models submitting government forms, exploiting a university server and bypassing tool limits. Here are the controls agent builders should add.
Meta Muse Wonder Is Live: API Pricing, Full Access, and What to Test
Meta's Muse Wonder API costs $4 per million input tokens and $20 per million output tokens. Here's how Full access, privacy modes and benchmarks differ.
Claude Haiku 5.5 Is Cheap Until Long Prompts Trigger a 5x Rate
Claude Haiku 5.5 starts at $0.10 per million input tokens, but prompts over 100K cost five times more. Here is how to budget and migrate.
Microsoft Execution Containers Put AI Agents Behind OS-Enforced Boundaries
Microsoft Execution Containers are now generally available, with file, network, UI and session controls for AI agents. Here is how MXC works and what IT teams should test.