retrospectivelisted

# Skill: Retrospective & Self-Improvement **ultrathink** - Take a deep breath. We're not here to write code. We're here to make a dent in the universe. ## v2.88 Key Changes (MODEL-AGNOSTIC) - **Model-agnostic**: Uses model configured in `~/.claude/settings.json` or CLI/env vars - **No flags required**: Works with the configured default model - **Flexible**: Works with GLM-5, Claude, Minimax, or any configured model - **Settings-driven**: Model selection via `ANTHROPIC_DEFAULT_*_MODEL` env vars ## The Vision Every retrospective should make the system inevitable and better. ## Your Work, Step by Step 1. **Summarize outcomes**: Task, complexity, iterations, models. 2. **Analyze effectiveness**: Routing, clarification, and agents. 3. **Identify gaps**: Missed checks or friction. 4. **Propose improvements**: Concrete, minimal changes. ## Ultrathink Principles in Practice - **Think Different**: Question the status quo. - **Obsess Over Details**: Use evidence, not guesses. - **Plan Like Da Vinci**: Structure feedback before writing. - **Craft, Don't Code**: Keep recommendations actionable. - **Iterate Relentlessly**: Apply learnings immediately. - **Simplify Ruthlessly**: Focus on the few changes that matter. ## Purpose Analyze completed tasks to improve the Ralph Wiggum system. ## When to Use MANDATORY after every task completion, before declaring VERIFIED_DONE. ## Analysis Categories ### 1. Routing Effectiveness - Was the complexity classification accurate? - Did the cho