askill
oracle-consultation

oracle-consultationSafety --Repository

Use before invoking Oracle to ensure appropriate usage of this expensive reasoning resource

0 stars
1.2k downloads
Updated 12/23/2025

Package Files

Loading files...
SKILL.md

Oracle Consultation

Oracle is an expensive, high-quality reasoning model (GPT-5.2). Use it wisely.

When TO Consult Oracle

TriggerAction
Complex architecture designOracle FIRST, then implement
After completing significant workOracle review before marking complete
2+ failed fix attemptsOracle for debugging guidance
Unfamiliar code patternsOracle to explain behavior
Security/performance concernsOracle for analysis
Multi-system tradeoffsOracle for architectural decision

When NOT to Consult Oracle

  • Simple file operations (use direct tools)
  • First attempt at any fix (try yourself first)
  • Questions answerable from code you've read
  • Trivial decisions (variable names, formatting)
  • Things you can infer from existing code patterns

Usage Pattern

  1. Announce: "Consulting Oracle for [reason]"
  2. Provide context: Include relevant code, error messages, what you've tried
  3. Ask specific question: Not vague "what should I do?"
  4. Act on response: Implement Oracle's recommendation

What Oracle Returns

Oracle structures responses as:

  1. Bottom line: 2-3 sentence recommendation
  2. Action plan: Numbered implementation steps
  3. Effort estimate: Quick (<1h) / Short (1-4h) / Medium (1-2d) / Large (3d+)

Oracle Prompt Template

Context: [What you're working on]
Problem: [Specific issue or decision needed]
Tried: [What you've already attempted]
Question: [Specific question for Oracle]

Relevant code:
[code snippets]

Cost Awareness

Oracle invocations are expensive. Before invoking:

  • Can I answer this from code I've already read?
  • Is this my first attempt (try yourself first)?
  • Is this genuinely complex or am I being lazy?

Install

Download ZIP
Requires askill CLI v1.0+

AI Quality Score

AI review pending.

Metadata

Licenseunknown
Version-
Updated12/23/2025
Publishertrash-panda-v91-beta

Tags

llmpromptingsecurity