How do I restore my saved context in a new session?

Once you start a new session, simply say 'start my session' or run the /nav:start command. The skill will detect the active marker and automatically offer to restore your state.

Does this skill automatically clear the Claude Code conversation?

No, due to technical limitations within Claude Code, the skill prepares the state marker and then provides instructions for the user to manually clear the context or start a new session.

When is the best time to use the nav-compact skill?

It is highly recommended once your conversation exceeds 50 messages, or immediately after completing an isolated feature to keep the AI's focus sharp.

How much token space does compacting actually save?

By summarizing the last 15-50 messages into a single structured marker, the skill typically achieves a 92% reduction in token usage for the next session.

Where are the context markers stored?

Markers are saved as markdown files within your project directory at .agent/.context-markers/, allowing them to be version controlled or shared.

Navigator Context Compactor

Name: Navigator Context Compactor
Author: alekspetrov

byalekspetrov

•

생산성 및 워크플로

Reduces Claude Code token usage by approximately 92% by creating persistent state snapshots before clearing conversation history.

소개

The Navigator Context Compactor skill optimizes long development sessions by solving the challenge of LLM context saturation. It acts like a 'git commit' for your AI conversation, capturing a comprehensive summary of technical decisions, modified files, and current progress into a localized context marker. By transforming a bloated message history into a concise snapshot, it allows developers to start fresh sessions with a 92% reduction in token usage while ensuring that the AI retains full awareness of the project's current state and next steps.

주요 기능

Estimated 92% token reduction for improved performance and lower costs
Message count analysis to determine the optimal time for compaction
Seamless auto-restoration workflow for starting new conversation threads
Automated context snapshots including technical decisions and file changes
47 GitHub stars
Intelligent session state persistence using the .active marker system

사용 사례

Resetting conversation clarity after completing a major sub-task or refactor
Maintaining project continuity when switching between complex feature branches
Optimizing performance during long-running sessions that approach context limits

소개

주요 기능

Estimated 92% token reduction for improved performance and lower costs
Message count analysis to determine the optimal time for compaction
Seamless auto-restoration workflow for starting new conversation threads
Automated context snapshots including technical decisions and file changes
47 GitHub stars
Intelligent session state persistence using the .active marker system

사용 사례

Resetting conversation clarity after completing a major sub-task or refactor
Maintaining project continuity when switching between complex feature branches
Optimizing performance during long-running sessions that approach context limits