The weekly review loop

Build-along · ~15 min at your own keyboard

A system without review drifts. Fifteen minutes weekly is enough. You ask three questions: what ran, what failed, and is there one improvement worth trying. At most one improvement, not five — small changes you can trace beat big rewrites you cannot. 'No change' is a good answer when the evidence shows it's working. Repairs are different: if something broke or did something unsafe, fix it now, however many that takes, and list each repair separately from the improvement. Log each review in a fixed template so week twelve is comparable to week one. Put the review on your calendar as a meeting with yourself.

Ladder · Level 6The long version: The weekly review ↗

Do this

0/6

Checkpoints

▸The review recurs on your calendar.
▸You have a filled-in entry for this week with evidence.
▸Repairs and the improvement (or 'none') are recorded separately.

The three ways this goes wrong

The review turns into rewriting everything. — Do the repairs, then stop at one improvement. Write the other ideas under 'Next' and move on.
Evidence says 'seemed fine'. — Open the run history. Paste a timestamp or a link. No timestamp, mark it unverified.
You skipped two weeks. — Cut the review to ten minutes and the template to three fields. Keep the slot. Missing the habit is a bigger risk than a thin review.

Dated facts in this lesson: Claude Cowork scheduled tasks ("Scheduled" in the sidebar) run a saved prompt in the cloud with the laptop off. They need Pro or above (researched Oct 4, 2026, not yet confirmed by a person: check the source before you rely on it); ChatGPT has agent, Projects (project-only memory, files, instructions), memory toggles in Settings > Personalization, and Scheduled Tasks at chatgpt.com/schedules. Workspace agents ("ChatGPT Work", July 2026) are on Business and Enterprise. Agent quotas: Plus 40 messages a month, Pro 400. Secondary source (researched Oct 4, 2026, not yet confirmed by a person: check the source before you rely on it); Gemini Spark is a 24/7 agent with Skills and Schedules. Schedules need Pro or Ultra, personal accounts only, up to 50 schedules. A Mac version exists. Spark is a beta (researched Oct 4, 2026, not yet confirmed by a person: check the source before you rely on it); Scheduling in personal Copilot is not confirmed. Power Automate is the Microsoft route: desktop flows are free with a Microsoft account, cloud flows come with Microsoft 365 (researched Oct 4, 2026, not yet confirmed by a person: check the source before you rely on it); Grok has Projects (files and instructions), "Personalize with Memories" under Data Controls, and Automations on a paid tier. SuperGrok Heavy ($300) runs parallel agents. Secondary source (researched Oct 4, 2026, not yet confirmed by a person: check the source before you rely on it); Perplexity has Memory (also on Free), Spaces (files and instructions; Pro allows 50 files per Space; not on Free), Tasks (at most hourly) and Computer on Max. Secondary source (researched Oct 4, 2026, not yet confirmed by a person: check the source before you rely on it).

Share what happened

Paste your review template and one filled-in review entry.

Strip API keys, passwords, and tokens before you paste. Swap them for [KEY].

🔒 Don't paste confidential, personal-health, or payment info.