Does this skill automatically execute the /compact command?

No, it only provides a suggestion. You retain full control over your session and decide whether or not to follow the prompt based on your current workflow.

Can I customize the frequency of the suggestions?

Yes, you can set the COMPACT_THRESHOLD environment variable to adjust the number of tool calls required before the first suggestion appears.

Why should I use manual compaction instead of automatic?

Automatic compaction triggers at arbitrary points, often purging crucial context in the middle of a complex task. Strategic Compact lets you choose logical breakpoints to ensure your implemention plan stays intact.

Is this skill specific to Japanese users?

While this specific documentation is provided in Japanese, the logic and shell script work globally for any Claude Code CLI user regardless of their primary language.

How does the skill track when to suggest a compaction?

The skill uses a PreToolUse hook to count the number of Edit and Write operations. Once you reach the threshold, it provides a non-intrusive suggestion to run the /compact command.

Strategic Compact

Name: Strategic Compact
Author: saar120

bysaar120

0•

생산성 및 워크플로

Optimizes Claude Code sessions by suggesting manual context compaction at logical task boundaries to prevent arbitrary data loss.

Strategic Compact is a performance-focused workflow enhancement for Claude Code designed to solve the problem of unpredictable automatic compaction. By monitoring tool usage via PreToolUse hooks, the skill suggests performing a manual '/compact' command at strategic intervals—such as after research phases or milestone completions—rather than allowing the system to trigger it during critical implementation steps. This ensures that essential context is retained when you need it most while keeping token usage efficient and sessions responsive during complex, multi-step development tasks.

주요 기능

01Integrates seamlessly with standard Claude Code setting hooks

02Prevents mid-task context loss from arbitrary auto-compaction

03Provides periodic reminders every 25 calls after the initial threshold

04Triggers alerts at configurable thresholds (default: 50 calls)

05Monitors Edit and Write tool calls to track session density

060 GitHub stars

사용 사례

01Clearing out heavy debugging logs and error context after successfully resolving a bug

02Managing long-running sessions that involve multiple distinct feature updates

03Switching from a deep research and planning phase to active code implementation

What are Skills?·How to Install

Install with 🐟 Skill.Fish

npx skillfish add saar120/skillgate-registry strategic-compact

For use in Claude.ai and ChatGPT

주요 기능

01Integrates seamlessly with standard Claude Code setting hooks

02Prevents mid-task context loss from arbitrary auto-compaction

03Provides periodic reminders every 25 calls after the initial threshold

04Triggers alerts at configurable thresholds (default: 50 calls)

05Monitors Edit and Write tool calls to track session density

060 GitHub stars

사용 사례

01Clearing out heavy debugging logs and error context after successfully resolving a bug

02Managing long-running sessions that involve multiple distinct feature updates

03Switching from a deep research and planning phase to active code implementation

What are Skills?·How to Install

Install with 🐟 Skill.Fish

npx skillfish add saar120/skillgate-registry strategic-compact

For use in Claude.ai and ChatGPT