Risk Kernel 0.1.1
RiskKernel is an open-source, self-hosted runtime that puts hard limits around AI agents: a per-run cost ceiling, a loop-iteration cap, a wall-clock budget, and a kill switch — all enforced in compiled Go, never by an LLM. You point an existing agent at it with one env var (OPENAI_BASE_URL), bring your own provider key, and nothing leaves your machine except the model call you were already making. It also does crash-resume, human approval gates on side-effecting tool calls, and OpenTelemetry export to whatever backend you already run. Apache-2.0, no telemetry.
| Tags | golang mcp reliability self-hosted sre observability ai-agents opentelemetry llms llmops agent-governance guardrails-ai |
|---|---|
| License | Apache |
| State | initial |
Homepage
Download
