operator-runtime
operator-runtime is a TypeScript SDK for building automation systems around an operator model.
operator-runtime is still in active development.
The core idea is simple:
- define the task as a
goal - pick a
capabilitythat names the task domain - pass structured
context - run with a real
runtime - let an
engineexecute - accept or reject the result with deterministic
verifylogic - inspect the
trace
Current
The current repo implements the operator runtime, capability registry, deterministic verification flow, trace recording, a browser-backed example, and a browser-backed integration test.
Planned
The docs discuss future runtime directions, including an LLM-backed runtime, but those parts are clearly marked as planned or WIP.
Why this exists
Most automation systems get muddy when task meaning, execution resources, and acceptance rules are mixed together.
This SDK separates them:
- the operator owns the run contract
- the capability names the task domain
- the engine performs execution
- the runtime supplies resources
- verification defines deterministic acceptance
Mental model
OPERATOR
│
│ task + context + verify
▼
Capability-selected execution
│
▼
Engine
│
│ uses runtime
▼
Runtime (Browser / APIs / Test Runtime)
│
▼
Environment
│
▼
Verification
│
▼
Result + Trace
Start here
- Read Architecture Overview for the full execution model.
- Read Capabilities, Context, and Runtime to understand the core boundaries.
- Follow Your First Operator to see the pieces together.