feat: Switch compaction model from session model

[ryanwyler]

Issue for this PR

Closes #13946
Closes #12135
Related: #14368, #14259, #14233

Type of change

• Bug fix
• New feature
• Refactor / code improvement
• Documentation

What does this PR do?

Adds a dedicated compaction model selector so users can run one model for chat and a different model for summarization. Run Claude Sonnet or Opus 4.6 for interactive coding while compacting with Zen Big Pickle (free) to eliminate compaction cost entirely, or GPT-5.3 Codex (400k context) for detail-oriented summaries that capture nuance your session model would lose — with headroom to compact sessions that have already exceeded their primary model's context limit.

For best results use this enhanced compaction prompt: #14662

Why not just use agent.compaction.model in config?

The existing agent.compaction.model config option is static — it requires editing opencode.json and is fixed for the lifetime of the config. This PR adds runtime switching without touching config files, matching how the primary model switcher works:

• Switch compaction models on the fly without restarting or editing config
• Different sessions can use different compaction models in the same instance
• Browse available models with the same familiar picker UI
• See the active compaction model in the prompt footer at a glance

Both approaches coexist. Model resolution priority: TUI selection > config file (agent.compaction.model) > session model. When no TUI selection is made, the config-based approach works exactly as before. Nothing is broken.

Features:

• Reuses existing provider infrastructure — same providers, API keys, and auth as the primary model picker
• /compaction-models slash command and "Switch compaction model" in the command menu
• Model resolution priority: compaction model override > agent config > session model
• Compaction model preference stored via kv.signal in kv.json, matching how all other TUI settings are persisted
• Prompt footer displays active compaction model when set
• When unset, behavior is identical to upstream

Changes:

• DialogModel accepts target="compaction" prop — no duplicate component
• SessionCompaction.process() accepts and resolves the override model
• CompactionPart schema extended with optional compactionModel field
• compaction_model_list keybind added (default: none)
• SDK regenerated via ./script/generate.ts

How did you verify your code works?

Tested with Claude Sonnet 4.5 (chat) + GPT-5.3 Codex (compaction). Confirmed correct model at each pipeline stage via log entries. Compaction completed successfully, session resumed on the original model. Also tested auto-compaction with no override set to verify identical upstream behavior. Verified that existing agent.compaction.model config continues to work when no TUI override is set.

Checklist

• I have tested my changes locally
• I have not included unrelated changes in this PR

---