TAG
A typed delegation protocol for routing work between agents, tools, models, and humans with scoped authority, structured handoffs, and replayable records.
Reliable agent systems need more than loose prompt chaining. TAG treats delegation as a governed protocol with explicit capability boundaries and evidence.
Problem Space
Multi-agent workflows can become ambiguous when tasks, permissions, outputs, and responsibilities are passed through unstructured prompts.
System Direction
TAG defines public-safe delegation concepts for capability classes, structured handoffs, scoped authority, and evidence-preserving work transfer.
Public Capabilities
- 01Typed delegation concepts
- 02Capability-bound handoffs
- 03Scoped authority framing
- 04Replayable work records
- 05Human-agent coordination patterns
TAG is presented publicly as a protocol direction. Internal schemas, enforcement mechanics, and routing algorithms are not disclosed.
What Is Not Disclosed
Private implementation details, security-sensitive internals, and unreleased runtime architecture are intentionally not disclosed.