skills.rest
JEROME-PRAKASH-L
JEROME-PRAKASH-LCOMMUNITY
Explore all skills by this author →
WebsiteGitHub

oracle

Bundles prompts and selected files for second-model code review via Oracle CLI.

1|Updated Jun 9, 2026
Search Tags:#code review#debugging#refactoring#second opinion#oracle cli#token estimation#gpt-5.5
One-click install
npx skills add https://github.com/JEROME-PRAKASH-L/openclaw --skill oracle-jerome-prakash-l
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: oracle
Source: https://github.com/JEROME-PRAKASH-L/openclaw/tree/main/skills/oracle
Command: npx skills add https://github.com/JEROME-PRAKASH-L/openclaw --skill oracle-jerome-prakash-l

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill requires @steipete/oracle.

What problem does it solve? Getting a second opinion on code from another model usually means manually copying files into a chat window, losing track of context, and wasting tokens. This Skill wraps the Oracle CLI to bundle a prompt plus a curated file set into one second-model pass for review, debugging, refactoring, or design feedback. ## Core Features & Use Cases - Dry-run previews: Check payload size and token spend with --dry-run summary --files-report before spending anything. - Two engines: Use the browser engine for long ChatGPT Pro runs, or the API engine when OPENAI_API_KEY or Azure config is available. - Session management: Long runs may detach; list sessions with oracle status and reattach with oracle session <id> --render instead of blindly re-running. - Use Case: You hit a stubborn bug in your TypeScript backend. You run a dry-run over src/** excluding tests, confirm the token cost, then launch a browser-engine run with a detailed prompt and reattach to the session when it finishes. ## Quick Start Ask the assistant to run an Oracle dry-run preview over your source directory with a description of the problem, then launch a browser-engine review run with the same file set.

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?▼

Use the Oracle CLI to bundle a prompt with selected files and send them to a second model in one pass. Run a dry-run first to preview the payload, then launch with the browser or API engine and treat the output as advisory.

How do I check token cost before running Oracle?▼

Run oracle with --dry-run summary --files-report plus your prompt and --file patterns. This previews the payload and token spend without consuming any tokens, so you can trim the file set before the real run.

Should I use the Oracle browser engine or API engine?▼

Use the browser engine for long ChatGPT Pro runs, which is the default when no API key is set. Use --engine api when OPENAI_API_KEY or Azure config is available, or when you need Claude, Grok, Codex, or multi-model runs.

What happens when an Oracle run times out or detaches?▼

Do not re-run it. Sessions are stored under ~/.oracle/sessions, so list recent runs with oracle status --hours 72 and reattach with oracle session <id> --render to retrieve the result.

How do I exclude test files from the Oracle file set?▼

Pass negated glob patterns as additional --file arguments, for example --file "src/**" --file "!src/**/*.test.ts". Oracle honors .gitignore, skips default-ignored directories like node_modules, and rejects files over 1 MB.

Related Skills

Explore More Software Engineering Skills→
Moliboy5000Moliboy5000

teach-impeccable

Explore codebases and ask UX, brand, and aesthetic questions to persist design principles.

Community
Intermediate
dilhanzdilhanz
9

ship-plan-verify

Verify implementation plans against codebase patterns and dependency checks.

Community
Advanced
gaebalaigaebalai
2

pr-quality-gate

Run automated quality checks on pull requests before submission.

Community
Intermediate
Z43LZ43L

claude-code

Run Claude Code CLI coding tasks in print or tmux interactive mode.

Community
Advanced