Fast non-interactive briefing before any meeting — auto-detects your next calendar event, pulls relationship history, surfaces open commitments, and produces a one-page brief in under 30 seconds. Use this whenever the user says "brief me", "give me a quick brief", "what's coming up", "background on my next call", "who am I meeting next", "brief me on Sarah", "I have a call in 10 min", "quick rundown", or right before walking into a meeting. Different from /minutes-prep — brief is the fast hook-fireable version that doesn't ask questions and doesn't set goals. Use brief when speed matters; use prep when the user wants to think hard about goals first.
Fast non-interactive briefing before any meeting — auto-detects your next calendar event, pulls relationship history, surfaces open commitments, and produces a one-page brief in under 30 seconds. Use this whenever the user says "brief me", "give me a quick brief", "what's coming up", "background on my next call", "who am I meeting next", "brief me on Sarah", "I have a call in 10 min", "quick rundown", or right before walking into a meeting. Different from /minutes-prep — brief is the fast hook-fireable version that doesn't ask questions and doesn't set goals. Use brief when speed matters; use prep when the user wants to think hard about goals first.
triggers
["brief me","give me a quick brief","what's coming up","background on my next call","who am I meeting next","brief me on Sarah","I have a call in 10 min","quick rundown"]
phase
lifecycle
user_invocable
true
metadata
{"display_name":"Minutes Brief","short_description":"Fast meeting briefing from your conversation memory.","default_prompt":"Use Minutes Brief to prepare me quickly for an upcoming meeting using my stored meeting history.","site_category":"Lifecycle","site_example":"/minutes-brief","site_best_for":"Get a fast one-page brief before an upcoming meeting."}
Fast, non-interactive briefing that synthesizes your relationship history with someone into a one-page brief — designed so you can read it in 60 seconds before walking into a call.
This is the proactive layer. It's built to be invoked silently by a hook (e.g., 15 min before a calendar event) and to also work as a manual /minutes-brief command. Unlike /minutes-prep, brief asks no questions and sets no goals — it just hands you the facts.
How it works
Phase 0: Determine the target
Three ways the user can invoke this:
1. With a name (/minutes-brief sarah, "brief me on Alex")
→ Use that name directly. Before searching, check for learned aliases:
If aliases exist, search across all returned variants and treat them as the same person for the rest of the flow. If the user explicitly says two names are the same person, persist it:
node "${CLAUDE_PLUGIN_ROOT}/hooks/lib/minutes-learn-cli.mjs" set-alias "Sarah Chen""Sarah""User confirmed these refer to the same person"
Then skip to Phase 1.
2. With "auto" or no argument (/minutes-brief, "brief me on my next call")
→ Auto-detect the next upcoming calendar event. Try sources in order — use the first that works:
Filter to events with 2+ attendees, skip all-day events, pick the soonest. Pull attendee names from the event.
gog CLI (if installed):
gog calendar list --today --json 2>/dev/null
Apple Calendar via osascript (every Mac, zero install):
osascript -e 'tell application "Calendar" to get {summary, start date} of (every event of every calendar whose start date >= (current date) and start date < ((current date) + 2 * hours))'
If none return anything, ask once: "I can't find an upcoming meeting. Who do you want a brief on?" Then take whatever they say and move on.
3. Hook-fireable mode (/minutes-brief --auto, or invoked silently from a hook)
→ Same as auto, but never ask questions. If no upcoming meeting and no name, exit silently with no output. Hooks should never spam the user.
Phase 1: Gather data in parallel
The Minutes CLI already does the hard work. Person profiles, search, relationship rankings, and commitments are bounded policy-safe projections. Run the commands below in parallel and require exit status 0 before interpreting their output.
# 1. Bounded live-source person profile.
minutes person "<name>"# 2. Recent decisions involving them, last 30 days.# Clean JSON on stdout by default (insights does NOT accept --json — output is already JSON).
minutes insights --participant "<name>" --kind decision --since <30-days-ago>
# 3. Recent meetings with them, last 60 days. Newline-delimited JSON (one object per line).
minutes search "<name>" --limit 10 --since <60-days-ago> --format json
CLI stream-handling notes — the Minutes CLI is actively developed and its stream contract is not fully settled. Today (0.8.0):
minutes person, minutes insights, and minutes search --format json are bounded live-source surfaces.
minutes people, minutes commitments, and minutes person fail closed on authorization, resource-budget, or correction races. Never suppress a nonzero exit or translate it into an empty fact.
Any nonzero exit from a live-source command means the source is unavailable. Do not interpret empty stdout as “no history.”
Do not invent new flags on top of what's shown above — e.g. minutes insights --json is not a real flag, minutes export --since is not a real flag. The CLI will reject unknown flags with a usage error.
If a future CLI release changes any of these contracts, update this skill in the same PR that ships the CLI change.
Before declaring "first meeting on record", verify it with live-source search:
minutes search "<name>" --limit 1 --format json
For multi-attendee meetings, focus on the requested person. Mention the others briefly at the end. Don't try to brief five people at once.
Phase 2: Read what you actually need
Extract the file paths of the most recent 1–2 meetings from a successful minutes search response. Reauthorize and retrieve each meeting through the native bounded surface:
minutes get "<exact path>" --json
Require exit status 0 and use only the returned content. Never pass the path to the host Read tool: search results are hints, not retained authorization capabilities.
Don't read more than two — past three meetings is enough context for a brief, and the brief is supposed to fit on one screen.
Phase 3: Synthesize the brief
Produce a brief in this exact shape — every section is one tight chunk, total fits on one screen:
# Brief: <PersonName> · <today's date>**Last conversation** (<dateofmostrecentmeeting>): <2–3 sentences in prose, including the emotional tone if discernible from the transcript>
**They've been thinking about**: <comma-separatedlistof3–5hottopicsfromthelast30days, orderedbyrecency × frequency>**Commitments**: Use `minutes commitments --person "<name>" --json`; require exit status 0.
**Where things stand**: <one-linereadoftherelationshipvibe — warming, cooling, stable, urgent, drifting>**Open with**: "<aconcretefirstsentencetheusercouldliterallysayatthetopofthecall — referencessomethingspecificfromthelastconversation>"
Concrete example of a good "Open with" line:
Bad: "Hey Sarah, how's everything going with Q2?" — generic, could be said to anyone.
Bad: "Hi Sarah, hope you're doing well!" — pure filler.
Good: "Hey Sarah — last time you mentioned the Q2 hiring freeze was making your roadmap impossible. Did legal come back on the contractor question?"
The good version is specific (cites Q2 freeze, the contractor angle), references something only THIS conversation history would surface, and gives Sarah an immediate hook to pick up the thread.
Rules for the brief:
No filler. Cut anything that doesn't change what the user does in the next 60 minutes.
Be specific. "They want to talk about Q2" is useless. "Sarah has raised the Q2 hiring freeze in 3 of her last 4 meetings" is useful.
Honest about gaps. If minutes search returns nothing, say that no matching meeting history was found. Never infer relationship facts from the unavailable graph.
The opening line matters most. This is the punchline. Make it usable verbatim. The user should be able to literally say it.
Phase 4: Save and display
Save the brief to ~/.minutes/briefs/ so /minutes-debrief can compare against it later:
…followed by the brief body. Then chmod 600 the file — briefs contain relationship intelligence and should be private.
Display the brief inline in the response. Don't paraphrase it or summarize it after — let the brief speak for itself.
Phase 5: One-line nudge
End with exactly one line — this skill is about speed:
"Want to think harder about goals? Run /minutes-prep <name>. After the call, run /minutes-debrief."
That's it. No follow-up questions, no "anything else I can help with?". Brief is fast on purpose.
Gotchas
Record explicit workflow preferences when the user states them. If the user says something like "default to prep", "always brief me first", or "stop reminding me about meeting prep", persist it:
Record explicit person aliases when the user confirms them. If the user says "Sarah and Sarah Chen are the same person" or "Dan means Dan Benamoz here", persist it:
Hook-fireable mode is silent on failure. When invoked with --auto and nothing matches, exit cleanly with zero output. Hooks should never spam the user. Errors only matter when the user is actively asking.
First-name slugs match the rest of the plugin. Save files as sarah.brief.md, not sarah-chen.brief.md. Slug rules: lowercase, ASCII-only (strip diacritics via Unicode NFKD), replace spaces and punctuation with hyphens, take only the first name token. "María José Pérez" → maria. "Jean-Claude" → jean-claude. Single-name people (Madonna) → madonna. This matches /minutes-prep and /minutes-debrief so they can find each other.
If the user types two names, ask once. "brief me on Sarah and Alex" — pick one as the focus and tell the user. "I'll focus the brief on Sarah since you have more history with her, and mention Alex briefly." Don't refuse, don't try to brief both fully.
Don't duplicate prep. Brief sets no goals, asks no questions, and never produces talking points. If the user wants any of that, the closing nudge points them at /minutes-prep. Two skills, two jobs — keep the line clean.
One screen of output. If the brief body is longer than ~25 lines, you've added filler. Cut it.
Multi-attendee meetings focus on one person. Pick the single most-mentioned attendee (highest meeting count) and brief on them. Mention others in one sentence at the end: "Also in the room: Logan, Kim." Don't try to brief everyone — the signal dilutes to noise.
Zero history is not an error — but verify it before claiming it. If minutes search returns empty, say that no matching meeting history was found and use only the calendar event. Never invent history.
Calendar attendee names are messy. The event might say "Sarah Chen sarah@acme.com" but transcripts say "Sarah" or "SPEAKER_1". Match on first name when looking up history; the CLI's --participant flag does fuzzy matching for you.
Briefs are sensitive. Always chmod 600. They contain relationship intelligence the user wouldn't want leaked.
Be honest about staleness. If the most recent meeting with this person is 3+ months old, lead with that fact: "Last conversation (4 months ago): …" — the user's mental model of the relationship may be more recent than reality.
Run the three live-source CLI calls in parallel. They're independent. Synthesize only from returned source-backed results after all three exit successfully.
A failed commitment projection is unavailable, not empty. Never claim “No open commitments” from a nonzero graph command.
