tkhribech/claude-founder-os / skills-library/weekly-review
weekly-review
VerifiedRun the founder's weekly business review — pull the week's numbers, compare against the running baseline, surface what changed and what needs a decision. Use when the founder says "weekly review", "how did the week go", "run the numbers", or at the first session of a new week.
Weekly Review
Turns "how did the week go?" into a five-minute, decision-ready briefing. Output is one dated markdown report plus updated pending tasks — not a data dump.
When to use / when NOT to use
- Use for the recurring weekly check-in, or when the founder asks how the business is doing.
- Do NOT use for deep-dive analysis of one channel (do that as a normal task) or for anything that changes live systems — this skill is read-only.
Inputs
- Last week's report: the newest
memory/status/weekly-review-*.md(the baseline). - The business's number sources, as defined in
memory/RULES.md§Review — e.g. for Kiln & Co: booking counts from the class system, shop orders, newsletter growth, social reach. Credentials come from the git-ignored credentials folder, read at runtime. memory/status/PENDING-TASKS.md(open loops to re-check).
Steps
- Load the baseline. Read the newest previous review. No baseline = say so and create the first one; never invent a comparison.
- Pull this week's numbers from each defined source. Record the number you actually saw, with its source — never an estimate presented as a reading.
- Compare and rank. For each metric: this week, last week, delta. Rank the report by decision-relevance, not by source order — the biggest mover leads.
- Sweep the open loops. Anything in PENDING-TASKS now due, unblocked, or stale gets a line.
- Draft the decision list. At most 3 items phrased as decisions the founder can make in one sentence ("extend early-bird a week: yes/no?").
- Verify before reporting: re-open the written report file; check every number in it against the value pulled in step 2; check the date math on every deadline.
Output
memory/status/weekly-review-YYYY-MM-DD.md — first line date-stamped, then: TL;DR (3 bullets) → metric table with deltas → open-loop sweep → decision list. In chat, lead with the TL;DR and the decision list; the file holds the detail.
Guardrails
- Read-only: no live-system writes, no sends, no schedule changes from inside this skill.
- Numbers without a named source don't go in the report.
- If a number looks implausibly good or bad, say so and check the source twice before reporting it (suspiciously clean results are broken verification until explained).
Related skills
| Skill | Stars | Updated | Works with | Status | Install |
|---|---|---|---|---|---|
| receiving-code-reviewobra/superpowersUse when receiving code review feedback, before implementing suggestions, especially if feedback seems unclear or technically questionable - requires technical rigor and verification, not performative agreement or blind implementation | Stars: 322k | Updated | Claude CodeCodex | Verified | Install |
| requesting-code-reviewobra/superpowersUse when completing tasks, implementing major features, or before merging to verify work meets requirements | Stars: 322k | Updated | Claude CodeCodex | Verified | Install |
| using-git-worktreesobra/superpowersUse when starting feature work that needs isolation from current workspace or before executing implementation plans - ensures an isolated workspace exists via native tools or git worktree fallback | Stars: 322k | Updated | Claude CodeCodex | Verified | Install |
| verification-before-completionobra/superpowersUse when about to claim work is complete, fixed, or passing, before committing or creating PRs - requires running verification commands and confirming output before making any success claims; evidence before assertions always | Stars: 322k | Updated | Claude CodeCodex | Verified | Install |