level-1-essentials/individual/09-rituals/weekly-monthly-rituals.md

Weekly and Monthly Rituals

The kit is not one-and-done. It needs a rhythm.

Weekly (Fridays, 30 minutes)

Monday morning (10 min)

• Skim release notes: OpenAI, Anthropic, Google Gemini, Perplexity
• Note anything worth trying this week
• Update model routing cheat sheet if a shift occurred

Wednesday afternoon (15 min)

• Try one model you don't normally use for a real task
• Notice what it does well or poorly
• Log to individual/05-model-routing/ if it changes your routing

Friday afternoon (30 min) — the big one

1. Run the weekly-review prompt (04-prompt-library/19-weekly-review.md)
2. Prune 3 things:
3. One memory entry that's stale
4. One prompt from your library that you didn't use
5. One subscription you're not using
6. Add:
7. New prompts to library based on real usage this week
8. New terms to your personal glossary
9. Check API spend on each provider dashboard
10. Confirm zero-retention still enabled on providers that support it

Monthly (last Friday, 60 minutes)

1. Memory audit — 06-memory-hygiene/memory-audit-template.md
2. Prompt library review — which prompts got used? Which didn't? Prune and refactor.
3. Model bake-off — pick one task you did 5+ times this month. Run it against the top 3 models. Update your cheat sheet if a challenger won.
4. Cost review — total AI spend, per provider. Any surprises?
5. Skill spike — pick one term from the glossary you don't feel fluent in. Spend 20 min going deep.

Quarterly (90 minutes)

1. Rotate API keys — all providers, all keys
2. Full glossary re-read — one hour
3. Rewrite Master System Prompt — from scratch, don't just edit. Compare to old version.
4. Automation audit — which of your automations still run? Any that need to be killed or rebuilt?
5. Read one deep paper or attend one AI conference talk
6. Re-do the fluency test in individual/08-jargon-glossary/learning-path.md

Annual (one afternoon)

1. Rebuild the eval set from scratch — new prompts, new expected outputs
2. Rewrite the personal style guide
3. Renew any DPAs / vendor contracts
4. Audit privacy line — has your risk tolerance changed?
5. Set next year's AI-skill goal (e.g., "learn to fine-tune", "build my first MCP server")

The calendar setup

Put these on repeat:

• Monday 8:30am — "AI release notes skim" (10 min)
• Wednesday 3pm — "AI experiment of the week" (15 min)
• Friday 3:30pm — "Weekly AI review" (30 min)
• Last Friday of month, 3pm — "Monthly AI audit" (60 min)
• First Monday of quarter, 2pm — "Quarterly AI audit" (90 min)
• Every 90 days — "Rotate API keys"

Anti-ritual: don't do these

• Don't chase every new model release — most won't matter to you
• Don't add to the prompt library without pruning
• Don't hoard subscriptions "in case I need them"
• Don't set-and-forget memory — audit it