oracle

Bundles user prompts with repository files for one-shot code analysis.

Updated Feb 8, 2026
One-click install
npx skills add https://github.com/nomad3/openclaw-k8s --skill oracle-nomad3
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: oracle
Source: https://github.com/nomad3/openclaw-k8s/tree/main/package/skills/oracle
Command: npx skills add https://github.com/nomad3/openclaw-k8s --skill oracle-nomad3

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

It reduces the back-and-forth of missing context by bundling your question with the exact repository files the model needs.

Core Features & Use Cases

  • One-shot repo-aware answers: Bundles your prompt with selected code/docs so another model can reason using real context (browser or API).
  • File selection patterns: Use globs, includes/excludes, and respects .gitignore to attach only the highest-signal files.
  • Sessions and reattachment: Supports long-running browser runs with stored sessions so you can reattach instead of re-running.

Quick Start

Run oracle in browser mode to analyze a task using the files under src while previewing what will be sent first: oracle --engine browser --dry-run summary -p "Diagnose and propose fixes for failing tests" --file "src/" --file "!/.test."

Frequently Asked Questions about oracle

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

FAQPage Schema
How do I bundle repository files with a prompt for one-shot code investigation?

To bundle repository files with a prompt for code investigation, specify file selection patterns like globs and includes to attach high-signal code directly to your question. This provides context-rich, one-shot answers by ensuring the model reasons using real repository context.

What is the best way to diagnose failing tests using repository context without back-and-forth?

Diagnosing failing tests using repository context is best achieved by bundling your diagnostic prompt with the exact source files needed. Use file globs to include relevant code and exclude tests, reducing missing context back-and-forth for a direct, one-shot analysis.

Can I preview token spend and file attachments before sending a large codebase to the model?

Yes, you can preview token spend and file attachments before sending a large codebase by using a dry-run preview. This feature safely handles large files by showing exactly what will be bundled, ensuring token spend awareness before generating the final prompt.

Does prompt bundling support reattaching to long-running browser sessions for codebase summaries?

Prompt bundling supports reattaching to long-running browser sessions for codebase summaries. It stores browser sessions so you can reattach to previous investigative contexts instead of re-running the entire file selection and bundling process from scratch.

How do I exclude files when attaching repository context to a prompt?

To exclude files when attaching repository context to a prompt, use negated file glob patterns alongside standard includes. The system respects .gitignore rules and applies deterministic file attachment rules, allowing you to attach only the highest-signal files for your investigation.