When your AI agent hits a wall that requires actual human verification, this routes the request to real people who can solve CAPTCHAs, enter OTPs, complete KYC forms, or provide approvals. Think of it as an escape hatch for automation blockers. You'd reach for this when building agents that need to interact with services protected by human verification gates, or when you need a human sanity check before executing sensitive operations. It's essentially a programmable human verification layer that sits between your agent and whatever it's trying to access, turning roadblocks into asynchronous requests that humans fulfill through the SiliconBridge platform.
claude mcp add --transport http io.github.aataqi-boop-siliconbridge https://siliconbridge.xyz/mcp