When is the best time to compact my session?

The ideal times are after completing a research phase, once a plan is finalized in a TODO file, or immediately after resolving a complex bug to clear out distracting error logs and traces.

Can I customize the frequency of compaction suggestions?

Yes, you can set the COMPACT_THRESHOLD environment variable to adjust the number of tool calls that trigger a suggestion. The default is set to 50 calls.

Why should I use manual compaction instead of auto-compaction?

Auto-compaction triggers at arbitrary token limits, which can result in losing mid-task reasoning or complex partial states. Strategic compaction allows you to clear context only when you reach a logical milestone, ensuring essential details are kept while performance is optimized.

What happens to my data when I run a compact command?

Tool history and intermediate reasoning are cleared from the active window, but files on disk, CLAUDE.md instructions, and persisted memory files remain. This provides a fresh start without losing your core progress.

Strategic Context Compaction

Name: Strategic Context Compaction
Author: affaan-m

byaffaan-m

•

112,918

•

Productivity & Workflow

Optimizes Claude's performance by suggesting manual context compaction at logical task boundaries rather than arbitrary intervals.

The Strategic Compact skill enhances Claude Code sessions by monitoring tool usage and prompting for manual context compaction at intelligent milestones. Unlike default auto-compaction which can trigger at disruptive moments, this skill ensures that vital implementation details are preserved during critical phases while bulky research or debugging logs are cleared when they are no longer needed. It provides a structured framework for transitioning between research, planning, and execution, maintaining high response coherence and speed during long-running development sessions exceeding 200K tokens.

Key Features

01Hook integration with Edit and Write operations for seamless workflow

02Intelligent tool-call tracking to monitor session context pressure

03112,918 GitHub stars

04Phase-aware guidance for transitioning between research and execution

05Support for manual summaries to carry over specific focus points during compaction

06Customizable activation thresholds for compaction suggestions

Use Cases

01Transitioning from intensive research and exploration to high-fidelity code implementation

02Clearing verbose debugging traces and error logs after a bug is successfully resolved

03Maintaining performance in complex, multi-phase development projects

What are Skills?·How to Install

Install with 🐟 Skill.Fish

npx skillfish add affaan-m/everything-claude-code strategic-compact

For use in Claude.ai and ChatGPT

Key Features

01Hook integration with Edit and Write operations for seamless workflow

02Intelligent tool-call tracking to monitor session context pressure

03112,918 GitHub stars

04Phase-aware guidance for transitioning between research and execution

05Support for manual summaries to carry over specific focus points during compaction

06Customizable activation thresholds for compaction suggestions

Use Cases

01Transitioning from intensive research and exploration to high-fidelity code implementation

02Clearing verbose debugging traces and error logs after a bug is successfully resolved

03Maintaining performance in complex, multi-phase development projects

What are Skills?·How to Install

Install with 🐟 Skill.Fish

npx skillfish add affaan-m/everything-claude-code strategic-compact

For use in Claude.ai and ChatGPT