oracle

Bundle task prompts with repository files for second-model code feedback.

28|Updated Jan 3, 2026
One-click install
npx skills add https://github.com/ma08/botfiles --skill oracle-ma08
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: oracle
Source: https://github.com/ma08/botfiles/tree/main/claude/skills/oracle
Command: npx skills add https://github.com/ma08/botfiles --skill oracle-ma08

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

Getting accurate, context-aware feedback from a second AI model on code tasks normally requires manually copying all relevant files and background context, which is slow, error-prone, and often misses critical project details.

Core Features & Use Cases

  • Prompt and File Bundling: Automatically packages your task prompt with selected repository files (supporting globs, exclusions, and .gitignore rules) to deliver full context to the second model.
  • Flexible Model Engine Support: Runs via default browser mode for ChatGPT GPT-5.5 Pro with manual login, or API mode for other models like Claude, Grok, or Codex.
  • Common Use Cases: Ideal for debugging stubborn issues, validating refactors, checking design decisions, and cross-verifying code changes against full project context.

Quick Start

Use the oracle skill to get a second-model review of your recent bug fix by bundling the relevant source files and your debugging prompt.

Frequently Asked Questions about oracle

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

FAQPage Schema
How do I get a second AI model to review my code using full repository context?

This approach automates bundling your prompt with relevant repository files, including globs and .gitignore rules, to provide full project context for a second-model code review.

What is the best way to validate a code refactor against project context?

Packaging updated source files with your prompt and sending the bundle to a second model cross-validates refactors against the broader repository context.

Does this second-model review approach work with API engines like Claude or Grok?

Yes, this second-model review approach supports API engines like Claude, Grok, or Codex, and also runs in default browser mode for ChatGPT GPT-5.5 Pro with manual login.

How do I manage token budgets when sending multiple repository files for a design review?

The system applies selective file attachment and token budget management rules to control usage costs when bundling repository context for a design review.

Why does manually copying files for AI debugging miss critical project details?

Manual file copying for AI debugging is slow and error-prone, whereas automated prompt and file bundling ensures full context-aware feedback without missing critical project details.