Adopt ManagedCode.Communication for runtime result and error contracts #23

Problem

The control-plane runtime needs explicit success and failure contracts across provider, orchestration, and policy boundaries.

Scope

• Track where ManagedCode.Communication should wrap provider, runtime, and orchestration flows
• Define the initial error categories and result shapes to standardize across subsystems

Out of scope

• General exception-policy changes unrelated to the control-plane runtime
• End-user copywriting for every error state

Implementation notes

• Keep the result pattern focused on public runtime boundaries
• Align result contracts with the domain model issue and future provider adapters
• Avoid inventing a parallel result abstraction

Definition of Done

• The issue identifies the first runtime boundaries that must use ManagedCode.Communication
• The planned result and problem categories are explicit enough to implement later without guesswork

Verification

• Review the issue body against the planned provider and runtime issues to confirm coverage

Dependencies

• Parent epic: [Epic] Establish dotPilot as a desktop control plane for local-first agent operations #11
• Depends on: Define the dotPilot domain model for agents, sessions, fleets, tools, artifacts, and evaluations