weekly-review

Most weekly reviews are diaries. Busy week, lots of progress. They change nothing. weekly-review takes a different stance: a review must close loops. You pull last week commitments from the prior log, reconcile them against what actually happened, and the discomfort it produces is exactly the signal that it is working. Sentiment ends here; what follows are three concrete artifacts: a reconciliation table, a carry-forward list with reasons, and a short set of checkable next-week goals.

The workflow runs five steps. Pull last week commitments from the prior log, not from memory. Mark each item done, partially done, or not done, and attach evidence to every completion claim. Record finished work as a fact — shipped, sent, merged — instead of a feeling. Carry every unfinished item forward with a fresh reason or drop it explicitly; silent carry-over is how dead tasks stay alive. Commit to a small set of specific, verifiable goals; a list of fifteen is next weeks failure pre-written. One dated log file, newest on top, means step one next week is a read, not an excavation. Across weeks the drift becomes visible: the same item carried forward three weeks running is no longer a plan, it is a decision waiting to be made.

It is for independent developers, researchers, freelancers, or anyone who notices the weekly check-in sliding toward self-congratulation. The first run has no prior log, so it simply sets a baseline. From the second week onward every review is a genuine calibration pass. Soft work — learning, relationships, mentorship — is genuinely hard to express as evidence; the skill prefers an honest admission of no clean evidence over a fabricated metric.