skills.rest
brikkoAI
brikkoAIOFFICIAL
Explore all skills by this author →
WebsiteGitHub

oracle

Bundle repository files and prompts into a single request for second AI models.

1|Updated May 3, 2026
Search Tags:#code review#debugging#AI workflow#oracle#second-model#file bundling#repo context
One-click install
npx skills add https://github.com/brikkoAI/brikko-studio --skill oracle-brikkoai
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: oracle
Source: https://github.com/brikkoAI/brikko-studio/tree/main/packages/core/skills/oracle
Command: npx skills add https://github.com/brikkoAI/brikko-studio --skill oracle-brikkoai

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

When you need a second AI model to debug, refactor, design, or review your code, it lacks critical context about your repository's structure, tech stack, and conventions, leading to generic or incorrect suggestions if you manually share partial information.

Core Features & Use Cases

  • One-shot context bundling: Combine your task prompt and selected repository files (with support for glob patterns and exclusions) into a single request for a second AI model, eliminating manual copy-pasting.
  • Flexible execution engines: Run via browser for GPT and Gemini models, or via API for Claude, Grok, and other models, with stored sessions you can reattach to if runs time out or detach.
  • Use case: Stuck on a persistent bug in your codebase? Use this skill to bundle the relevant source files, error logs, and your debugging prompt to get a detailed, context-aware analysis from a second model.

Quick Start

Use the oracle skill to bundle your code review prompt and the relevant source files from your repository to send to a second AI model for context-aware feedback.

Frequently Asked Questions about oracle

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I provide full repository context to a second AI model for code review?▼

Bundle your code review prompt and selected repository files into a single request using glob inclusion and exclusion patterns, eliminating manual copy-pasting. This provides the second AI model with full repository context for context-aware feedback.

Can I use file glob patterns to select specific source files for AI debugging context?▼

Yes, the skill supports file glob inclusion and exclusion patterns to select specific source files for AI debugging context. This allows you to precisely control which repository files are bundled into the prompt sent to the second model.

Does this approach work with both browser-based and API execution engines for different AI models?▼

Yes, it supports browser execution for GPT and Gemini models, and API execution for Claude, Grok, and other models. Stored sessions allow you to reattach to long-running runs if they time out or detach.

What is the best way to optimize token spend when sending repo context to an external AI model?▼

Use the token spend preview feature to optimize context usage before sending repo context to an external AI model. This helps manage costs and ensures the bundled prompt stays within the second model's context limits.

Why does my second AI model give generic or incorrect suggestions when refactoring my codebase?▼

A second AI model gives generic or incorrect suggestions when refactoring your codebase because it lacks critical context about your repository's structure, tech stack, and conventions, often caused by manually sharing partial information.

Related Skills

Explore More Software Engineering Skills→
rohunvorarohunvora
5

prompt-engineering

Design prompts and templates for LLM interactions and Agent commands.

Community
Advanced
batFormatbatFormat

aif-evolve

Analyze project context and patches to improve AI Factory skills.

Community
Advanced
act70255act70255

code-reviewer

Automate code reviews with static analysis and AI vulnerability detection.

Community
Advanced
MOODMNKY-LLCMOODMNKY-LLC

oracle

Run targeted code review tasks with token-efficient dry-run previews.

Community
Advanced