docs: improve README self-upgrade/model-switching + multi-session brainstorming roadmap

[chinkan]

What changed

1. Improved README sections

• Moved Self-Upgrade, Model Switching, and Soul Files from orphaned table rows into a proper "Agent Lifecycle & Runtime" subsection under Features
• Rewrote descriptions with user-centric language (why you'd use each feature, not just what the command is)
• Added relevant emoji per section

2. Multi-Session & Multi-Model brainstorming

• Added a "Planning phase" placeholder section in README
• Created docs/roadmap/multi-session-multi-model.md with:
  • Motivation: concurrent sessions with independent models
  • Core concepts: sessions, sessions, model binding, per-session tools
  • Proposed /session slash commands
  • Design decisions to explore (session isolation, context sharing, concurrency)
  • Open questions (storage, model cold start, UX)

Before / After

Before:

---
| 🚀 Self-Upgrade | /self-upgrade slash command + self_upgrade tool... |
| 🔀 Model Switching | /models slash command to switch OpenRouter LLM models... |
| 📝 Soul Files | SOUL.md / AGENTS.md / USER.md persistent identity files... |

After:

### Agent Lifecycle & Runtime

| Capability | Description |
| Self-Upgrade | Trigger an in-place upgrade: pulls from git source or downloads the latest GitHub release binary. Auto-restarts after upgrade — no SSH, no manual steps. |
| Model Switching | Switch OpenRouter models at runtime via /models. Interactive picker lets you choose the best model per task: fast/cheap for simple queries, powerful for complex reasoning. |
| Soul Files | SOUL.md (persona), AGENTS.md (behaviour), USER.md (preferences) — persistent identity files auto-injected into every system prompt. Session-end self-reflection with .bak backups. |

Next steps

• Review and discuss the multi-session design
• Merge this PR (docs-only, no code changes)
• Create follow-up PRs for implementation