SuperAPI AI Directory

oil-gas-delegation

General & other

View repository

Updated 10-16 06:22

About this skill

Skill: Oil & Gas Delegation

Meta-skill that detects software tasks and routes them to appropriate petropowers skills.

Purpose

Oil & gas domain skills handle interpretation, analysis, and optimization recommendations. When a request requires building software (web apps, dashboards, APIs, databases), this skill delegates to petropowers skills.

Detection Logic

When a request contains any of these keywords, delegate to software development workflow:

  • Web app, dashboard, visualization
  • API, REST endpoint, GraphQL
  • Database schema, data model, persistence
  • Data pipeline, ETL, automation script
  • Report generation system
  • Alert notification system

Decision Flow

User request received
  ↓
Does request require building software?
  ↓
  YES → Invoke oil-gas-delegation
        ↓
        Route to appropriate petropowers workflow
  ↓
  NO → Handle within domain skill

Delegation Workflows

Web Application / Dashboard

Request: "Build a drilling monitoring dashboard"

  1. Invoke petropowers:brainstorming skill

    • Refine dashboard requirements
    • Identify data sources, users, features
    • Create design document
  2. Get user approval on design

  3. Invoke petropowers:writing-plans skill

    • Break into implementation tasks
    • Define file structure, tech stack
  4. Execute via petropowers:subagent-driven-development

    • TDD for components
    • Integration tests
    • Code review between tasks

Data Pipeline

Request: "Create an automated LAS file processor"

  1. Invoke petropowers:brainstorming
  2. Get design approval
  3. Invoke petropowers:writing-plans
  4. Execute via petropowers:subagent-driven-development

API Development

Request: "Build a REST API for production data"

  1. Invoke petropowers:brainstorming
  2. Get design approval
  3. Invoke petropowers:writing-plans
  4. Execute via petropowers:subagent-driven-development

Quick Script

Request: "Write a script to convert SEG-Y to numpy"

  1. Check if truly needs TDD (one-off vs. production)
  2. If production: Use TDD workflow
  3. If one-off: Direct implementation with verification

Domain-Only Tasks (No Delegation)

These tasks stay within domain skills:

  • "Analyze this seismic section for faults"
  • "Interpret well log data for porosity"
  • "Calculate probability of success for this prospect"
  • "Optimize drilling parameters"
  • "Forecast production decline"
  • "Assess pipeline integrity"

Domain skills provide analysis, recommendations, and domain expertise. Software skills build tools to support that work.

Integration Points

Domain skills invoke this skill when they detect software requirements:

## Software Tasks

If the user requests software development:

1. Acknowledge the request
2. Invoke `petropowers:oil-gas-delegation`
3. Delegation skill routes to appropriate workflow
4. Return to domain work after software is complete

Example Interaction

User: "I need a dashboard to visualize my seismic data"

Agent response:

  1. Recognize: "dashboard" → software task
  2. Invoke: petropowers:oil-gas-delegation
  3. Delegation skill:
    • Invokes petropowers:brainstorming for dashboard design
    • Design: "Seismic Visualization Dashboard"
      • Upload SEG-Y files
      • Display inline/crossline/time slices
      • Filter by amplitude range
    • User approves design
    • Invokes petropowers:writing-plans for implementation plan
    • Executes via TDD workflow

User: "What's the porosity in this well based on the logs?"

Agent response:

  1. Recognize: Domain analysis, no software
  2. Invoke: petropowers:oil-gas-pipelines/exploration
  3. Domain skill:
    • Reads LAS file
    • Calculates porosity from density/neutron logs
    • Returns analysis result

Usage Instructions

When implementing this skill:

  1. Check user request for software keywords
  2. If software: Route to brainstorming workflow
  3. If domain: Stay in domain skill
  4. After software complete, return to domain context

This skill is the bridge between oil & gas domain expertise and software engineering best practices.