Human-Agent Collaboration Protocol (HACP)

A framework for collaborative work between humans and intelligent agents. Modular, open, and built for trust.

Core Principles

Hybrid Work as Default

Work is shared across humans and agents based on strengths.

Roles as Functions, Not Titles

Dynamic roles adapt to the loop, not the org chart.

Flow Over Sprints

Loops emerge responsively via real-world signals.

Transparency by Design

Every action is logged and attributable.

Cognitive Load Awareness

Participants define how much attention they can give.

Trustworthy Autonomy

Agents must explain their actions to gain trust.

Roles

Initiator (Human)

Starts new loops from insight or needs.

Decider (Human)

Owns scope-specific decisions.

Synthesizer (Human)

Integrates inputs into meaning.

Validator (Human)

Reviews quality or value of outputs.

Narrator (Human)

Explains process and context.

Signal Scanner (Agent)

Detects when loops need to begin.

Orchestrator (Agent)

Coordinates human and agent roles.

Implementer (Agent)

Executes routines and tasks.

Explainer (Agent)

Justifies decisions and reasoning.