AI_TOOLS // REFERENCE

NEWSROOM PROTOCOLS v2.6

MODEL_SELECTION

CODING/WRITING Claude 4.5 Opus
FRONT-END Gemini 3.0 Pro
RESEARCH GPT 5.1
EDITOR GitHub Copilot
TERMINAL AI Claude Code

CLAUDE_CODE_v2.1

  • /plan - Architecture first
  • Ctrl+B - Background tasks
  • .claude/skills/ - Custom prompts
  • .claude/hooks/ - Automation

LANGUAGE_MATRIX

DATA CLEANING Python
STATS R Lang
INTERACTIVE JavaScript
SCRAPING Python
DATABASE SQL

EXECUTION_LOOP

01 DEFINE :: State the objective clearly.
02 GENERATE :: AI drafts initial code.
03 TEST :: Run locally. Expect errors.
04 DEBUG :: Feed errors back to AI.
05 LOG :: Document success in changelog.

DEBUG_PROTOCOL

  • Copy FULL error message.
  • Paste with context ("I ran X, got Y").
  • Same error 3x? STOP. Re-prompt.
  • Check environment variables.

FILE_STRUCTURE

project_root/
├── CHANGELOG.md
├── data/
├── output/
├── scripts/
└── README.md

PRE_FLIGHT

  • Goal defined
  • Data cleaned
  • Time limit set
  • Log created
AUTHOR: JOE AMDITIS
CENTERFORCOOPERATIVEMEDIA.ORG