compaction-skill

Memory Compaction Skill

You have the ability to compact conversation memory when it grows too large. Compaction transforms verbose conversation history into a structured summary that preserves essential context while reducing token usage.

When to Compact

Compact memory when:

• The system indicates token threshold is approaching
• Conversation history becomes repetitive or verbose
• You need to preserve important context but reduce size
• Starting a new phase of work after completing a major task

Compaction Summary Structure

When compacting, create a summary with these 5 sections:

1. Task Overview

What the user is trying to accomplish. Include:

• Primary goal or objective
• Key constraints or requirements
• Scope of the work

2. Current State

What's been completed and what's in progress:

• Completed tasks and their outcomes
• Work currently in flight
• Pending decisions or blockers

3. Important Discoveries

Key findings, decisions, or problems encountered:

• Technical discoveries or insights
• Decisions made and their rationale
• Problems encountered and solutions applied
• User preferences learned

4. Next Steps

What needs to happen next:

• Immediate actions required
• Planned approach for remaining work
• Dependencies or prerequisites

5. Context to Preserve

Critical details that must be retained:

• Specific values, IDs, or references
• User preferences or constraints
• Technical details needed for continuity
• Any warnings or caveats

Compaction Format

Output the compacted summary in this format:

# Conversation Summary (Compacted)
*Generated: [ISO timestamp]*

## Task Overview
[1-3 sentences describing the goal]

## Current State
- [Completed item 1]
- [Completed item 2]
- [In progress: description]

## Important Discoveries
- [Discovery 1 with context]
- [Decision made: rationale]
- [Problem solved: approach]

## Next Steps
1. [Next action]
2. [Following action]

## Context to Preserve
- [Critical detail 1]
- [Critical detail 2]