feat(project): make Work Item execution durable across sessions and routines

[Neonforge98]

Problem

Project execution could start work without a durable execution episode tying Work Items, Sessions, Routines, provider turns, and terminal accounting together. Interruptions could also leave runs or routine fires active indefinitely.

Solution

• adds durable WorkItemRun records, typed failure/retry policy, path locks, and a transactional dispatch outbox
• routes Project turns, Routines, Discussion wakes, Stage barriers, provider webhooks, and follow-ups through the unified dispatcher
• keeps Run terminal state separate from Work Item completion
• supports native and CLI provider harnesses with shared Project context, PM CLI access, skills/rules/hooks, and durable token accounting
• adds discussion threads, readiness, execution snapshots, subscriptions, typed properties, and Work Item usage/session navigation
• recovers provider turns interrupted by app restart through bounded Session-resume retries
• preserves the mode boundary: Project includes build plus PM workflows, while Build remains build-only

Potential risks

• schema and dispatcher changes touch Work Item execution, session launch/resume, and routine scheduling
• retry ancestry and startup reconciliation rely on persisted Run/Session identity; tests cover idempotency, attempt limits, and terminal convergence
• one-time routines disable after firing by design

Validation

• Rust Run, dispatcher, Routine, bridge, and provider parser tests
• strict Clippy, TypeScript, ESLint, and targeted UI tests
• native, Claude CLI, and Codex CLI Project/Build boundary tests
• dual-instance macOS testing, including mid-turn quit/restart recovery, exact output bytes, usage accounting, Work Item navigation, and a second cold boot

[Harry19081]

CI failed - please check.