Where are the full responses stored?

All raw data is saved to a structured directory at /tmp/aura-frog/ on your local machine, organized by type and timestamp for easy reference.

Can I still access the full data if the summary is insufficient?

Yes. The skill includes patterns for referencing full data, allowing you to run a command to read the specific temporary file whenever granular detail is needed.

Does it automatically clean up old files?

The skill includes recommended cleanup commands to purge the temp directory daily or manually, ensuring your local storage doesn't grow indefinitely.

How does Response Analyzer save tokens?

It writes large outputs, such as 500-line test logs, to a local file and only provides Claude with a 10-line summary, preventing thousands of unnecessary tokens from cluttering the context window.

Response Analyzer

Name: Response Analyzer
Author: nguyenthienthanh

bynguyenthienthanh

•

생산성 및 워크플로

Optimizes Claude Code performance by offloading large command outputs to temporary storage and providing concise context summaries.

The Response Analyzer skill for Claude Code is a powerful utility designed to prevent context window saturation during complex development tasks. By intelligently redirecting verbose command results, massive API responses, and long file lists to local temporary files, it ensures that only high-level summaries and actionable insights are injected back into the conversation. This architecture allows developers to maintain long-running sessions without hitting token limits, achieving up to 96% token savings while keeping granular data accessible on-demand for deeper debugging when necessary.

주요 기능

01Structured Temp Directory: Organizes outputs into dedicated folders for commands, APIs, and search results.

021 GitHub stars

03Smart Summarization: Automatically extracts key fields from JSON, test failures from logs, and top results from searches.

04On-Demand Retrieval: Allows the AI to reference full datasets via specific IDs only when granular detail is required.

05Context Bloat Reduction: Saves significant tokens by offloading verbose data to local storage.

06Pattern-Based Triggers: Optimized specifically for build logs, test suites, and large JSON payloads.

사용 사례

01Processing large JSON API responses to extract specific data points efficiently.

02Managing extensive file search results during large-scale repository refactoring.

03Analyzing massive test suite failures without exceeding the AI context limit.

What are Skills?·How to Install

Install with 🐟 Skill.Fish

npx skillfish add nguyenthienthanh/aura-frog response-analyzer

For use in Claude.ai and ChatGPT

Download Skill