Sandboxing
3 videos across 3 channels
Sandboxing enables running code and AI workflows inside isolated, secure environments so organizations can experiment and automate without risking core systems. The videos survey architectures for sandboxed agents, secure harnesses and memory/planning features, and dynamic isolated workers that spawn on demand, with practical demos of agent platforms, AI gateways, and code-based tool use in secure sandboxes.
Ship your first Managed Agent
The video demonstrates how to ship your first production ready incident response agent using Claude manage agents, cover
Agentic AI with Vercel: Saltbox One
A Saltbox team member walks through how they use Verscell to build a Salesforce focused agent platform, detailing archit
Give Your LLM a Sandbox: Introducing Dynamic Isolated Workers
The video introduces MCP code mode, a reliable, code-based alternative to traditional multi-step tool calls for LLMs, im