メニュー
Generate the Weekly Dispatch -- a full week retrospective covering wins, commitments, open items, team signals, decisions, red flags, and next week preview.
Codex または Claude でインストール この Prompt をコピーして Codex、Claude、または他のアシスタントに貼り付けると、Skill ページを確認してインストールできます。
SOC 職業分類に基づく
| name | weekly-dispatch |
| description | Generate the Weekly Dispatch -- a full week retrospective covering wins, commitments, open items, team signals, decisions, red flags, and next week preview. |
Generate the Weekly Dispatch -- a comprehensive week-in-review that covers what happened, what's stuck, and what's coming. Published to the Briefing Book every Friday by 8 PM.
Collect data from all available sources for the current week (Monday through today). Run these in parallel where possible.
Read all memory/sessions/*-summary.md files from this week. These contain decisions, commitments, tension maps, and user state observations.
pm_list filter "open" -- all open items, check which are overduepm_list filter "done" -- items completed this week (look at completed_at dates)Read data/o365/calendar/YYYY-MM/ and data/google/calendar/YYYY-MM/ for this week and next week. Count meetings per day.
Search recent email for threads involving leadership or VIPs that indicate strategic shifts, budget discussions, org changes, or escalations. Use memory_search with date filters.
Search data/o365/teams/ for this week's messages. Look for patterns: who's communicating what, escalations, blockers.
Read data/fireflies/transcripts/YYYY-MM/ for this week. Identify key meetings and extract decisions/commitments.
Search data/limitless/lifelogs/YYYY-MM/ for this week. Look for off-calendar conversations, impromptu decisions, emotional signals.
Search data/imessage/daily/ for this week. Look for personal context items (family, health, mood).
Organize findings into these sections. Be specific. Use names, dates, numbers. No vague language.
What went right this week. Ship dates hit, deals closed, hires made, blockers cleared, processes improved. Be generous but honest -- if something was a win, say so. If it was mediocre, don't list it.
Two tables:
Pull from PM items. Cross-reference with session summaries and meeting transcripts for status updates.
Items that were open at the start of the week and still are. Focus on the ones that matter -- stale blockers, delayed decisions, aging promises.
What's happening with the people. Read between the lines of meetings, emails, and messages. Who's stepping up, who's struggling, what dynamics are shifting. Be observant but not gossipy. The user uses this to calibrate 1:1s and interventions.
Decisions made or forced this week. What was decided, by whom, and what it means. Include decisions the user made and decisions that were made around them.
Things that could blow up if ignored. Be direct. These are the items where the observation-not-instruction principle matters most -- make the implication visible, let the user decide what to do.
Calendar overview for next week. Key meetings, deadlines, travel. What prep is needed. What's likely to be contentious.
Quick metrics table: PM open/overdue/completed, meetings this week, Qdrant vectors, key hires pending, days since [relevant milestone].
Write to: ~/Edwin/briefing-book/docs/1. Briefs/Weekly Dispatch -- YYYY-WNN.md
Where WNN is the ISO week number.
Format: Follow the existing Weekly Dispatch format exactly (see prior dispatches in the Briefs folder). Use the same section headers.
After writing, publish to Obsidian:
python3 briefing-book/scripts/obsidian-publish --all
SKILL_COMPLETE: weekly-dispatch
STATUS: success | error
ARTIFACT: ~/Edwin/briefing-book/docs/1. Briefs/Weekly Dispatch -- YYYY-WNN.md
WEEK: YYYY-WNN
WINS: [count]
RED_FLAGS: [count]
PM_COMPLETED: [count this week]
PM_OVERDUE: [count]
ERRORS: [any errors, or "none"]
Meeting-to-closure decision tracking -- harvest decisions actually made in the last week's meetings (the weekly priorities forum first), maintain a persistent ledger, verify downstream follow-through evidence, and surface quietly-dying decisions at the two-week mark.
Monthly red-team pass over the user's active big bets -- build the strongest evidence-backed case AGAINST each position they're invested in (steelman opposition, not strawman compliance), ending each bet with an honest KILL / HEDGE / PROCEED-EYES-OPEN verdict.
Weekly knowledge-graph curation loop -- freshness sweep against the week's reality, stale-edge invalidation, pending-merge queue drain, and org-chart temporal completeness against your authority model.
PM Weekly Triage
Decision radar + dossier generator -- detect decisions approaching the user in the next few days and build evidence-backed dossiers (prior positions, stakeholder map, constraining commitments, what it gates) before the moment arrives.
Draft-first email posture -- high-priority emails and key-contact unreplied threads arrive each morning WITH a reply already drafted in the user's voice. The user edits-and-sends or discards instead of starting cold. Drafts only, never sends.