What is the primary benefit of the Oracle skill?

Oracle allows you to bundle large sets of files and a prompt into a 'one-shot' request, providing deep repository context to high-reasoning AI models.

Does Oracle support local .gitignore files?

Yes, Oracle automatically honors .gitignore rules when expanding file globs to prevent sensitive or unnecessary files from being uploaded.

Can I preview what data is being sent to the AI?

Yes, you can use the --dry-run flag along with --files-report to see exactly which files are included and estimate the token spend before making a request.

What happens if a long-running request detaches?

Oracle stores sessions locally. If a session detaches or the CLI times out, you can use the 'oracle session ' command to reattach and retrieve the output.

How does the browser engine work in Oracle?

The browser engine uses automation to send requests to web-based AI interfaces, which is ideal for long-running 'long-think' models that might timeout over standard APIs.

Oracle Context Bundler

Name: Oracle Context Bundler
Author: ishuru

byishuru

0•

ブラウザ自動化

Bundles prompts and source files into high-context requests for advanced AI reasoning via browser automation or API.

Oracle is a specialized CLI utility designed to bridge the gap between large local codebases and high-reasoning AI models like GPT-5.2 Pro. It automates the process of gathering relevant files, applying glob patterns, and bundling them into a single 'one-shot' request, allowing the AI to provide answers grounded in the actual project context. With unique features like session persistence, browser automation for 'long-thinking' tasks, and detailed token spend reports, Oracle is essential for developers performing deep architectural reviews or complex debugging that exceeds standard chat context limits.

主な機能

01Smart file bundling with support for complex glob patterns and .gitignore logic.

02Automatic filtering of large files, dotfiles, and common build directories.

03Integrated browser automation for utilizing long-running, high-reasoning models.

04Dry-run mode with token estimation and payload previews to manage costs.

050 GitHub stars

06Persistent session management to reattach to and track long-running AI tasks.

ユースケース

01Conducting deep-dive architectural analysis on large, multi-file repositories.

02Automating complex debugging sessions that require extensive codebase context.

03Generating comprehensive documentation and patch plans for legacy systems.

What are Skills?·How to Install

Install with 🐟 Skill.Fish

npx skillfish add ishuru/ishuru oracle

For use in Claude.ai and ChatGPT

Download Skill