technical-strategy

This skill guides you to make architectural and design decisions that stand the test of time and scale appropriately.

Core Principles

1. Your Creations Require Maintenance:

   • Cost of Ownership: Every line of code written is a line of code that must be debugged, upgraded, and secured.
   • Minimize Surface Area: Prefer solutions that require less code and fewer moving parts.
   • Lifecycle Management: Plan for deprecation from day one. How will this be replaced?

2. Don't Optimize the Wrong Problem:

   • Premature Optimization: Don't optimize for massive scale if you don't have users yet.
   • Frequency Matters: Don't spend days automating a task that happens once a year.
   • Good Enough Performance: If a query takes 200ms and runs once a day, don't spend a week making it 10ms.

3. You Don't Have to Do What Google Does:

   • Context is King: Solutions that work for 27,000 engineers (monorepos, custom build systems) may crush a team of 5.
   • Right-Size Solutions: Choose tools appropriate for the current team size and maturity.
   • Evaluate Trade-offs: Understand why big tech uses a tool before adopting it.

Actionable Instructions for Agents

• When choosing a technology or pattern:

  1. Read Context: Check docs/company_context.md. If missing, run .agent/workflows/interview_company_context.md.
  2. Assess the maintenance burden. Who will patch this?
  3. Check if the solution is appropriate for the scale. Is it too complex for the current problem size?
  4. Avoid "hype-driven development". Stick to boring, proven technologies unless the new tech solves a critical, specific pain point.

• When planning deployment or operations:

  1. Read Context: Check docs/operational_context.md. If missing, run .agent/workflows/interview_operational_context.md.
  2. Ensure decisions align with Ownership, Budget, and Release constraints.
  3. Identify where secrets and artifacts should live.

• When optimizing code:

  1. Ask "Is this a bottleneck?"
  2. If no data suggests it's a problem, leave it alone.
  3. Focus on readability and maintainability over clever optimization.