genAI2026

Read MEMORY.md at session start and apply [LEARN:tag] corrections; when the user corrects you, append [LEARN:category] wrong → right to MEMORY.md. Use when starting work on this project or when the user gives a correction to remember.

Run holistic pedagogical review on lecture slides. Checks 13 patterns: narrative arc, notation density, pacing, motivation before formalism, worked examples, fragment reveals, box hierarchy. Use when the user asks for pedagogy or teaching-quality review of slides.

Run proofreading on lecture files (.tex or .qmd). Checks grammar, typos, overflow, consistency, and academic quality. Produces a report only; does not edit files. Use when the user asks to proofread slides or check grammar/typos in lecture content.

Adversarial QA comparing Quarto slides to Beamer benchmark. Run critic (find issues) then fixer (apply fixes); repeat until approved or max 5 rounds. Use when the user asks to QA Quarto vs Beamer or validate a lecture translation.

Review R scripts for code quality, reproducibility, and domain correctness. Checks set.seed, paths, structure, figure quality, and conventions. Use when the user asks to review R code or check scripts in scripts/R or Figures.

Run combined multi-dimensional slide review: visual audit, pedagogy review, and proofreading (and optionally TikZ parity). Use when the user asks for full or comprehensive slide review before a milestone.

Perform visual layout audit of Quarto or Beamer slides. Checks overflow, font consistency, box fatigue, spacing, and layout. Use when the user asks for a visual or layout audit of slides.

Compile a Beamer LaTeX slide deck with XeLaTeX (3 passes + bibtex). Use when compiling lecture slides.

Show current context status and session health. Use to check how much context has been used, whether auto-compact is approaching, and what state will be preserved.

Create new Beamer lecture from papers and materials. Guided workflow with notation consistency.

Challenge slide design with 5-7 pedagogical questions. Checks ordering, prerequisites, and cognitive load.

Extract reusable knowledge from the current session into a persistent skill. Use when you discover something non-obvious, create a workaround, or develop a multi-step workflow that future sessions would benefit from.

Installs Ollama, sets up local models, and deploys a custom system prompt. Invoke when user wants offline/local LLM setup or to run a model with a specific prompt.