search

TOOLING-ITD-004: AI Tooling

hashtag
CONTEXT

Multiple AI tools exist for documentation: Cursor, Claude, ChatGPT, Copilot, etc. Teams have preferences. Mandating a single tool creates friction.

hashtag
PROBLEM

Should we mandate specific AI tools for documentation?

hashtag
OPTIONS CONSIDERED

  1. Mandate Cursor + Claude - Standardize on what works

  2. ✅ No mandate, document what works - Share recommendations, allow flexibility

  3. Ban AI tools - Ensure "authentic" human writing

hashtag
REASONING

Option 1 (Mandate): Creates friction for team members with different preferences. Tools evolve fast—today's best choice may not be tomorrow's.

Option 2 (Document what works): Share what the framework authors use (Cursor + Claude) as a reference. Let teams choose. Outcomes matter, not tools.

Option 3 (Ban): Absurd. We're an AI-forward company.

hashtag
IMPLICATIONS

  • No required AI tooling

  • Reference setup documented: Cursor IDE + Claude for iteration

  • Teams free to use alternatives (ChatGPT, Copilot, etc.)

  • Focus on output quality, not tool choice

Current recommendation (January 2026): Claude Opus 4.5 is the preferred model for spec work—significantly better reasoning and technical depth than alternatives. This will change; update this ITD when it does.

PreviousTOOLING-ITD-003: Custom Lint RulesNextTOOLING-ITD-005: API Documentation

Last updated