skills.rest
joelklabo
joelklaboCOMMUNITY
Explore all skills by this author →
Website𝕏@joelklaboGitHub

oracle

Orchestrate four AI agents to analyze engineering tasks and synthesize a report.

Updated Dec 31, 2025
Search Tags:#automation#workflow#cli#multi-agent#decomposition#oracle#synthesis
One-click install
npx skills add https://github.com/joelklabo/markdowntown-cli --skill oracle-joelklabo
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: oracle
Source: https://github.com/joelklabo/markdowntown-cli/tree/main/codex/skills/oracle
Command: npx skills add https://github.com/joelklabo/markdowntown-cli --skill oracle-joelklabo

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill requires gemini, copilot, claude, codex, patch, git, and includes scripts (resource) components.

What problem does it solve?

Oracle coordinates a four-agent critique and synthesis loop to produce a single, authoritative analysis of code, plans, or architecture, reducing ambiguity and speeding decision-making.

Core Features & Use Cases

  • Multi-model critique: runs parallel analysis across Gemini, Copilot, Claude, and Codex to surface diverse perspectives.
  • Structured decomposition: breaks tasks into four expert domains and synthesizes a prioritized Master Report.
  • Context-aware analysis: optionally injects repository workspace context to ground insights in actual code and files.
  • CLI automation: provides scripts to drive analysis from the command line and enable reproducible workflows.

Quick Start

Run the Oracle CLI with a prompt or file to generate a comprehensive intelligence report.

Frequently Asked Questions about oracle

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

FAQPage Schema
How do I run multi-model AI analysis on my codebase?▼

Parallel multi-model analysis decomposes complex engineering tasks across four expert domains, using Gemini, Copilot, Claude, and Codex to critique outputs and synthesize a prioritized Master Report. This reduces ambiguity and speeds up decision-making for code and architecture.

How do I synthesize multiple AI agent outputs into a single report?▼

Multi-model critique works by running parallel analysis across Gemini, Copilot, Claude, and Codex to surface diverse perspectives on engineering tasks. The outputs are then synthesized into a single prioritized Master Report to provide an authoritative analysis.

Do I need specific AI tools configured to run multi-agent code analysis?▼

Yes, to run multi-agent CLI automation for code analysis, you need Gemini, Copilot, Claude, and Codex configured with their CLI integrations. The skill requires these dependencies to orchestrate reasoning, critique, and synthesis across the four agents.

Can I inject repository context into multi-agent AI analysis?▼

Yes, you can optionally inject repository workspace context to ground insights in actual code and files. This context-aware analysis ensures the parallel agents synthesize a report based on your specific local workspace.

What is the best way to decompose complex engineering tasks for AI review?▼

The best way to decompose complex engineering tasks is using a structured breakdown into four expert domains, orchestrated by the Oracle CLI. It runs parallel analysis across four models and synthesizes a prioritized Master Report for fast decision-making.

What are the limitations of using a single-model fast path for code analysis?▼

The fast path for single-model code analysis provides results quickly but bypasses the multi-model critique loop. It skips the diverse perspectives gained from running Gemini, Copilot, Claude, and Codex in parallel, potentially increasing ambiguity in the final synthesis.

Related Skills

Explore More Software Engineering Skills→
slpy22slpy22

crosscheck

Query Gemini, Codex, Copilot, and Claude in parallel and synthesize responses.

Community
Advanced
itallstartedwithaideaitallstartedwithaidea
31

brainstorming

Refine vague software feature requests into an implementation-ready design brief.

Community
Intermediate
gooftroopgooftroop

git-worktree-agent-workflow

Enforce isolated git worktree workflows for agent code modifications.

Community
Advanced
hainamchunghainamchung
54

agent-assistant-code-fast

Execute coding tasks immediately by bypassing the planning phase.

Community
Intermediate