feat: Collapse Compaction - preserves recent context while summarizing older messages

[ryanwyler]

Summary

Adds a new "collapse" compaction mode that preserves recent context while summarizing older messages, with intelligent merging of historical summaries across multiple compaction cycles.

The problem: Standard compaction compresses the entire conversation into a tiny summary, causing massive context loss (100% -> 3%). The AI forgets everything. Additionally, when multiple compactions occur, each new summary only knows about messages since the last compaction - historical context from earlier summaries is permanently lost.

The solution: Collapse compaction addresses both problems:

1. Selective compression - Only compresses OLD, stale messages while keeping recent work pristine:

   • Extracts oldest 65% of tokens and summarizes them at a breakpoint
   • Uses newest 15% as reference context for the summary LLM
   • Leaves 35% of recent conversation completely untouched
   • Result: Compresses to 15-25% usage, leaving 75% context window available

2. Historical summary merging - When generating a new summary, includes up to N previous compaction summaries in the prompt, instructing the LLM to merge and consolidate all historical context. This creates a "rolling summary" that carries forward critical knowledge across unlimited compaction cycles:

   • Working code blocks preserved verbatim
   • User directives and preferences maintained
   • Design decisions and lessons learned accumulated
   • No information loss regardless of session length

Changes (v1.1 - Modular Rebase)

This version refactors the collapse logic into a self-contained module for easier future rebasing:

File	Lines	Purpose
compaction-collapse.ts	+530	New - All collapse logic isolated here
compaction.ts	+20	Minimal router to collapse module
config.ts	+36	Config schema options
id/id.ts	+63	insert() function for breakpoint IDs
message-v2.ts	+25	Logging in filterCompacted
index.tsx	+13	TUI toggle command
sidebar.tsx	+6	Show compaction method in sidebar
types.gen.ts	+24	SDK types (auto-generated)

Key architectural decision: Bulk of feature (~530 lines) is in compaction-collapse.ts with minimal touchpoints in existing files. This makes future rebases much easier - just bolt the module back on.

Configuration

{
  "compaction": {
    "method": "collapse",
    "trigger": 0.85,
    "extractRatio": 0.65,
    "recentRatio": 0.15,
    "summaryMaxTokens": 10000,
    "previousSummaries": 3
  }
}

Option	Default	Description
method	"standard"	Compaction method: "standard" or "collapse"
trigger	0.85	Trigger compaction at this fraction of context
extractRatio	0.65	Fraction of oldest tokens to extract and summarize
recentRatio	0.15	Fraction of newest tokens to use as reference
summaryMaxTokens	10000	Target token count for summary output
previousSummaries	3	Number of previous summaries to include for merging

Default method remains "standard" - users can opt-in to collapse via config or TUI toggle (Ctrl+P -> "Use collapse compaction").

Related

• Original PR: feat: Collapse Compaction maintains agent intelligence beyond compaction by preserving recent interactions and curating knowledge collapsing the older stale messages anomalyco/opencode#6374
• Queued request fix: separate PR at fix/queued-request-processing

Testing

1. Set compaction.method to "collapse" in config or toggle via TUI
2. Have a long conversation that triggers context overflow
3. Verify collapse message appears at breakpoint position
4. Verify recent messages remain untouched
5. Verify no infinite compaction loops
6. Fork a compacted session and verify it works correctly
7. Trigger multiple compactions and verify previous summaries are merged into new summary