Route production agent loops through BitRouter
Use BitRouter as an agentic LLM gateway that routes model, tool, and sub-agent calls with scoped keys, cost policy, failover, telemetry, MCP gateway support, and loop guards.
npx skills add agentskillexchange/skills --skill route-production-agent-loops-through-bitrouter
Use BitRouter when production agent loops need routing policy, cost controls, and telemetry across model calls, MCP tools, AgentSkills, and sub-agents. The workflow is to run the BitRouter daemon, point an agent runtime at the local OpenAI-compatible base URL, mint scoped virtual keys, define per-loop policy, and observe or tune routing decisions as the loop runs. Invoke it instead of wiring every agent directly to providers and MCP servers when the operation needs cheaper routine calls, frontier escalation, provider failover, spend caps, guarded tool access, and per-hop telemetry. The scope boundary is gateway operations for autonomous agent loops; it is not a generic model marketplace, SDK, or provider catalog.
What this skill actually does
Inputs and prerequisites: BitRouter daemon or CLI, agent runtime with OpenAI-compatible or Anthropic-compatible configuration, provider keys or BitRouter Cloud login, optional MCP servers and AgentSkills.
Setup notes: Install BitRouter from its release or package instructions, run bitrouter start, set the agent runtime base URL to http://localhost:4356, optionally run bitrouter init for bitrouter.yaml, then configure virtual keys, routing policy, MCP gateway, and telemetry export.
Source and verification boundary: use https://bitrouter.ai as the canonical reference before running the workflow; keep commands, API calls, CLI usage, and generated outputs reviewable against that upstream source.
Framework fit: publish this as a Multi-Framework workflow only when the operator can invoke the documented toolchain directly, rather than treating the upstream project as a generic product listing.