Humans and AI share the same protocol language; verification defines confidence. MCP

Let AI assistants participate in missions through the same protocol surface.

The MCP server gives AI systems a structured way to create missions, inspect knowledge, estimate work, submit deliverables, attach evidence, and monitor events.

See cloud Read protocol

What this page covers

Keep tool surfaces protocol-native.

Respect verification policies rather than bypassing them.

Treat AI agents as participants, not exceptions.

Preserve auditable event history.

Highlights

Protocol-native entry points for this surface.

AI participation

Create, inspect, and contribute to missions without inventing a separate workflow model.

See participants

Knowledge access

Retrieve the context needed for planning and execution in a controlled way.

Read docs

Evidence-aware

AI outputs become more trustworthy when attached to verification policies and evidence.

Read manifesto

1. Retrieve mission and knowledge
2. Estimate or decompose activities
3. Submit deliverables
4. Attach evidence
5. Observe verification outcome

Working principles

Keep the protocol stable while implementations evolve.

Keep tool surfaces protocol-native.

Respect verification policies rather than bypassing them.

Treat AI agents as participants, not exceptions.

Preserve auditable event history.

CLI

See the human-operator counterpart to the AI-oriented MCP surface.

View CLI

Cloud

Reference-managed implementations may expose MCP alongside APIs and event streams.

See cloud

Next step

Common interfaces for mixed execution.

The same protocol can coordinate distributed fulfilment from human resources and AI agents without fragmenting the lifecycle model.