Weekly review
A Friday review of the week — wins, open loops, stale tasks, and setup for next week. Activate when the user says "/assistant:report", "weekly review", "weekly report", "wrap up my week", or on a Friday-afternoon schedule.From its SKILL.md
npx -y skills add daddia/my-assistant --skill weekly-reviewAssembled from the repository path, not quoted from the project. Check it against their README if it does not work.
One thing to look at
- 1 stars1 stars. Stars are a popularity signal and not a quality one, but at this level it is likely that nobody has read this closely except its author, and you would be relying on your own review.
SKILL.md
3.5 KB, as published. Nobody here has run it
Weekly review
Close the week cleanly: what got done, what's still open, what's gone stale, and what to line up for Monday. Also the moment to prune tasks and memory. Catalog entry: job_id: weekly-review in scheduled/schedules.yaml.
Read the profile first
Load goals/priorities (to judge what mattered), voice, and working hours.
Schedule health check (optional, interactive Friday runs)
On an interactive Friday run after the expected window (cron 16:00 Friday + 90 minutes = 17:30), optionally surface a low-priority miss hint for weekly-review when:
scheduled/weekly-review.yamlexists andsurfaceislocalreview-{today}.mdis absent andlast_run_atis stale
Use the same ### Schedule health block format as daily-brief, naming the weekly review job. Suggest cloud-code (no managed cookbook for this job). Skip when surface is managed or cloud-code. Max one health block per chat turn; do not block the review.
What it does
- Wins — what closed this week. Pull from
TASKS.mdDone, and from the week's briefs/follow-ups if present. - Open loops — To do, In progress, In review, and Blocked tasks plus
follow-up-trackingitems still outstanding, most-important first. - Stale — tasks untouched 14+ days and follow-ups long cold. Present each for triage: still relevant? in progress? blocked? cancelled?
- Memory prune — surface stale or contradictory memory entries to confirm or remove.
- Next week — what to line up for Monday: upcoming meetings needing prep (hand to
meeting-prep), deadlines from goals, top 3 priorities.
Output shape
Weekly review — week ending Fri 11 Jul
Wins (6 closed)
• Shipped Q3 scope with Acme • Expense report in • ...
Open loops (4)
• Q3 SOW draft — due next week
• Acme contract — nudged twice, still silent (escalate?)
Stale — triage these
• "Fix back fence" — untouched 21 days → cancel or keep in To do?
• "Call accountant" — no context, 16 days → still needed?
Next week
• Mon 9:00 board prep (deck ready?) • Renewal deadline Thu
• Top 3: SOW, board deck, hiring loop
Scheduled use
Runs well as a Friday 4pm scheduled task; save to review-YYYY-MM-DD.md. See skills/schedule-setup/SKILL.md.
Heartbeat: at end of a scheduled run, update the ledger via bash (required):
python3 scripts/update_ledger.py \
--job-id weekly-review \
--status {success|partial|failed} \
--artifact review-{today}.md \
--working-folder {working-folder}
Before ending, confirm last_run_at in scheduled/weekly-review.yaml matches this run.
Rules
- Triage interactively — don't auto-drop or auto-reschedule tasks.
- Update
TASKS.mdandmemory/only with confirmation (writes to those are allowed, but a review is a good moment to confirm). - Read-only on email/calendar; draft-only on any nudge.
Approval frame
Follow rules/approval-frame.md when surfacing recommendations that need user decisions.
Use all four sections for actionable recommendations. Do not write queue items for read-only review sections (closed items, stale task lists) unless a separate pending approval artefact exists — e.g. a nudge draft or profile diff proposal.