agentleFS
derrekyoungderrekyoung/ai-chief-of-staff/.claude/skills/daily-briefing/SKILL.md

daily-briefing

Skill2 starsChanged 48 days agoMITView on GitHub

Give the user a two to three minute morning brief built from their tracker, yesterday's note, and the last week. Covers what slipped, today's three priorities, open loops, who is waiting on them, and what is coming this week. Use when the user says "brief me", "good morning", "what does my day look like", "what should I focus on today", or runs the briefing directly. Prints the brief in chat and publishes it to the Morning Brief artifact page. Its only file write is applying taps from the artifact pages to the tracker.

36 other projects use a version of this.

Adopted by

Most recently taken up by

Used with

Projects with it also use crm-lookup, spotify and 10 more.

Other files people use alongside this one.

Files found in the same public projects as a version of this one, more often than they turn up anywhere else. Each project counts once, large collections count for less, and this file's own repository is left out.
  • QuixiAISkill

    crm-lookup

    Search, view, update, and merge contacts in the local CRM

    Also in 3 projects that use this

  • QuixiAISkill

    spotify

    Search Spotify, inspect current playback and devices, and control playback with explicit approval

    Also in 3 projects that use this

  • MicrosoftSkill

    inbox-triage

    Use this skill whenever the user asks to clean up, triage, declutter, or archive low-value Outlook inbox mail - "clean up my inbox", "triage my mail", "declutter my inbox", "sort out newsletters", "archive old resolved threads", "/inbox-triage" - to group only newsletters, auto-notifications, past-event meeting logistics, resolved threads, and redundant duplicates into reviewable move-proposals presented for per-bucket approval before any message moves. Do not use for reading mail, drafting or sending replies, deleting mail, executing moves without explicit per-bucket approval, or acting on messages from the user's manager, direct reports, active threads, flagged mail, or labelled/HR/Legal/Finance/Security senders.

    Also in 4 projects that use this

  • MicrosoftSkill

    image-gen

    Also in 3 projects that use this

  • AnthropicSkill

    weather

    The optional standing status report ("the weather"): compile open incidents, build health, merge-queue stats, and deploy lag into one always-current report page, and post to the channel only when a defined event fires. Use when the weather routine fires, or when someone asks to run/update the weather report.

    Also in 3 projects that use this

  • SentrySkill

    notion

    Search, fetch, create, update, and move Notion pages and data sources. Use when users ask to look up docs, specs, notes, meeting notes, project context, roadmaps, trackers, or internal references stored in Notion, or when they explicitly ask to create or edit Notion content.

    Also in 3 projects that use this

  • +2 more from QuixiAI

The file

Show its rules and text

We split this skill into the rules it gives an agent

34 rules. 0 are common, 24 are distinctive, and 10 are very specific to this project.

  • 0

    Common

    Most of the 33 edited versions we read keep these.

  • 24

    Distinctive

    This author's own choices. Worth a look.

  • 10

    Very specific

    Tied to their project's names, files or tools. Probably skip.

  • Common
  • Distinctive
  • Very specific
  1. Very specific:

    Run "Applying taps" in artifacts/README.md first.

    line 20
  2. Distinctive:

    If it changed the tracker, commit that on its own, by path:

    line 25
  3. Distinctive:

    If a source is missing or empty, skip it.

    line 61
  4. Very specific:

    Read the timezone out of config/profile.md first, then attach it to every date command.

    line 66
  5. Distinctive:

    Never read a weekday off an ISO string.

    line 75
  6. Distinctive:

    If the profile has no timezone in it, stop and ask for one.

    line 82
  7. Distinctive:

    Use that same zone for every time you state.

    line 88
  8. Very specific:

    Read items only from the thread sections below the ## Sections heading.

    line 93
  9. Very specific:

    Read config/profile.md and config/season.md before you write a word.

    line 123
  10. Distinctive:

    Use the season to decide what to emphasize, what to downplay, and what counts as relevant this morning.

    line 127
  11. Distinctive:

    Check the answer slots, not the file text.

    line 135
  12. Distinctive:

    Every slot still unfilled, nothing written at all: the user has not been set up.

    line 148
  13. Very specific:

    If reviews/ holds a weekly review dated yesterday or today, open the brief with one line pointing at it, for example "Your weekly recap is …

    line 164
  14. Distinctive:

    Do not keep a separate list of things not to mention.

    line 215
  15. Distinctive:

    Only when a calendar connector is available: Check for one before you write this section.

    line 239
  16. Distinctive:

    When a calendar is live, pull today's events in the user's timezone and lay them out simply: time, what it is, who is in it.

    line 244
  17. Distinctive:

    If more than three feel urgent, the season breaks the tie.

    line 259
  18. Distinctive:

    Always a bullet list. Tracker items due in the next seven days that are not already above, plus unresolved threads from recent daily notes …

    line 265
  19. Distinctive:

    If there are more than six, group the tail under one "also on the list" line.

    line 276
  20. Distinctive:

    Only include what is genuinely significant or needs a head start today.

    line 284
  21. Distinctive:

    Only when there is something real.

    line 289
  22. Very specific:

    Check patterns/ first. If the theme already has a file, say it is continuing rather than presenting it as a discovery.

    line 297
  23. Distinctive:

    If an email connector is live, this is where a thread waiting on the user's reply belongs.

    line 311
  24. Very specific:

    No Artifact or ArtifactData tool in the session: skip the page silently.

    line 331
  25. Very specific:

    Add timezone, ask_context, and the calendar fields.

    line 336
  26. Very specific:

    Write it to tmp/brief.json with the Write tool, then ArtifactData set, collection: "briefs", doc_id = today's date, file_path = …

    line 337
  27. Very specific:

    Do not write to daily/ and do not write to reviews/.

    line 344
  28. Distinctive:

    Write out abbreviations, "Monday" not "Mon"

    line 362
  29. Distinctive:

    Use digits for numbers, "3" not "three"

    line 363
  30. Distinctive:

    Prefer bullets, short sentences, and natural connective phrases

    line 364
  31. Distinctive:

    No special characters or symbols

    line 365
  32. Distinctive:

    No em dashes and no double hyphens anywhere, use commas, parentheses, or a new sentence

    line 366
  33. Distinctive:

    No spoken headers, transitions between sections should sound natural rather than announced

    line 368
  34. Distinctive:

    Avoid parenthetical asides that read badly out loud

    line 370

“Common” is measured: kept by at least 6 in 10 of the 33 edited versions of this file we could read. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.

Keep what fits and drop the rest. Your version, ready for your agents.
Files
SKILL.md

Daily Briefing

Opens the day. Reads the tracker, yesterday's note, and the last week, then hands back a short, honest read on what is slipping, what matters today, and who is waiting on you.

The brief prints in chat and lands on the Morning Brief artifact page, where the user can tick off or snooze tracker items, see today's calendar live, and ask Claude about any item. It writes no daily note and no review. The one file it touches is tracker.md, and only to apply taps the user already made on the artifact pages.

Before anything else: apply page taps

Run "Applying taps" in artifacts/README.md first. The brief reads the tracker, so yesterday's taps have to be in it before you read it. When both URLs in config/artifacts.md are blank, or there is no ArtifactData tool in the session, this step does nothing and you say nothing about it.

If it changed the tracker, commit that on its own, by path:

git add 'tracker.md'
git commit -m "tracker: apply page taps 2026-03-16"

Unmatched taps get one line each at the very top of the brief, before everything else. Applied taps are not worth mentioning.

When this runs

In the morning, on request. The user asks to be briefed, says good morning, asks what the day looks like, or runs the skill by name.

Nothing needs to exist for it to work. If yesterday has no note, say that in one line and keep going. If the tracker is empty, the brief gets shorter, not padded.

Sources (read in this order)

  1. tracker.md. The to-do list. It drives Top of Mind Today, Today's Priorities, Open Loops, and Coming Up This Week. Read it before anything else.
  2. Yesterday's daily note, daily/YYYY-MM-DD.md. Look for the End of Day Reflection and the ## Decisions Made section the wrap writes, spelled exactly that way.
  3. The last seven daily notes. Texture, unresolved threads, anything repeating.
  4. Recent files in meetings/, roughly the last two weeks. What was agreed, and what has not moved since.
  5. patterns/. Existing themes, so a Pattern Alert continues a known thread instead of announcing it as new.
  6. people/, for anyone you are about to name.
  7. Calendar, only if a connector is live. See Today's Schedule below.

If a source is missing or empty, skip it. Do not invent content to fill a section, and do not narrate the absence more than once.

Dates

Read the timezone out of config/profile.md first, then attach it to every date command. A bare date reads the machine clock, which is a different day from the user's whenever they are travelling or the machine is set to something else.

With Europe/Lisbon in the profile:

  • TZ=Europe/Lisbon date +%Y-%m-%d for today's date
  • TZ=Europe/Lisbon date +%A for the weekday

Never read a weekday off an ISO string. For past days, compute them the same way with the zone still attached. Past date arithmetic is not portable: macOS has BSD date and takes TZ=Europe/Lisbon date -v-1d +%A, Linux has GNU date and takes TZ=Europe/Lisbon date -d yesterday +%A. Run one, and if the shell rejects the flag, run the other. You can also copy the weekday out of the daily note's own heading.

If the profile has no timezone in it, stop and ask for one. Do not guess a zone, do not use the machine zone, and do not brief around the gap. Every date in the brief depends on it, and a brief built on the wrong day is worse than a brief that waited thirty seconds for an answer. Setup asks for the timezone and does not let it be skipped, so this should never come up.

Use that same zone for every time you state. If a day you want to reference cannot be pinned down, say "a few days ago" rather than guessing a weekday.

Reading the tracker

Read items only from the thread sections below the ## Sections heading. Everything above it documents the format, and one line there has the shape of an unchecked item without being one. Never scan the whole file for - [ ], and skip any line holding a placeholder in curly braces. On a fresh tracker that rule is the difference between an empty list, which is the truth, and a phantom task called {item text}.

Parse every item for three markers:

  • priority: !high, !med, !low
  • due date: due:YYYY-MM-DD
  • capture date: <!-- captured: YYYY-MM-DD -->

Then bucket each open - [ ] item:

  • Overdue. due: is before today.
  • Due today. due: is today.
  • Due this week. due: falls after today and within seven days.
  • High priority, no date. !high with no due:.
  • Everything else. Including undated, unprioritized items.

Skip - [x] items and the Completed (Archive) section, unless something closed yesterday is worth a line in Yesterday in 30 Seconds.

The captured: date matters here too. An undated item that has been sitting for six weeks is a candidate for What's Slipping even though nothing is technically overdue.

Current season

Read config/profile.md and config/season.md before you write a word. The profile gives you the user's name, timezone, and life threads. The season gives you the lens.

Use the season to decide what to emphasize, what to downplay, and what counts as relevant this morning. When two things compete for a priority slot, the one the season points at wins. If config/season.md names things to track without being asked, check for them in the last week's notes even when the user has not mentioned them.

When the config is not fully filled

Check the answer slots, not the file text. Both config files carry instruction comments that name the placeholder markers in prose, and those comments stay there forever, so a plain search for that string matches on a fully configured vault every time. Look at what sits under each heading instead. Nothing in config/templates/ is config: those slots are filled by skills on every run, so never read them as setup answers.

A slot counts as filled when it holds anything the user put there, and (skipped) is something the user put there. Treat it as a filled answer, a deliberate pass on an optional question, never as a blocker.

Then degrade rather than stop:

  • Every slot still unfilled, nothing written at all: the user has not been set up. Say so in one line, offer the setup interview in CLAUDE.md, and stop.
  • Some filled, some not: brief on what is there. No season means no season lens, so the brief runs without one and says nothing about it. No life threads means the tracker reads as one flat list.

The one gap you cannot work around is the timezone, so ask for it and wait, as the Dates section says.

A partly filled profile is a normal state, not an error.

What to include

Generate these in order. If a section has nothing real to say, leave it out. A short brief is a good brief.

If reviews/ holds a weekly review dated yesterday or today, open the brief with one line pointing at it, for example "Your weekly recap is ready: reviews/2026-03-15-weekly.md." When recaps_url in config/artifacts.md is set, point at the page instead: "Your weekly recap is ready on the Recaps page." That is how a recap written overnight by a scheduled run gets noticed. Do not summarize it; the user will read it.

Yesterday in 30 Seconds

Three to five bullets on what the user did, decided, or felt yesterday. Write it like a friend catching you up, not a log replay. No timestamps. If the wrap ran, lean on the reflection and the decisions. If yesterday's note is missing, one line saying so, then move on.

Top of Mind Today

From the tracker, in this order:

  1. Overdue items, prefixed with OVERDUE.
  2. Items due today.
  3. High priority items with no due date that are not already listed above.

Cap it at five. If there are more, name the count and surface the ones that carry the most weight. If all three buckets are empty, skip the section.

What's Slipping

One to three things quietly falling off. There is no gap report to read, so scan for it yourself across the last seven daily notes and the tracker:

  • Action items written into a daily note that never made it to the tracker. A line ending in see [[tracker]] is capture's pointer to an item that did make it, so it never counts. To decide whether a note item reached the tracker, compare the two the way capture's dedupe does, and ignore anything the user tacked on afterward ("done 3/27", "sent it"), since an annotated copy of a tracker item is still that item.
  • Tracker items whose captured: date is more than three weeks old with no movement in the notes.
  • A thread from a meeting note with an owner and no follow up.
  • Someone the user said they would contact, in the notes, with no sign they did.
  • Anything the season says to track that has not appeared in a week.

Curate hard. Three items land, ten items get ignored. Say what is slipping and since when, not what to do about it.

A thread is closed once any later note says it is. "Called Nadia back" on Wednesday closes Monday's "need to call Nadia", so before you flag something, read forward from where it appeared. If the user replies to the brief that something is already handled ("did that yesterday", "stop flagging it"), that reply is a capture: file it through the capture skill as a daily entry, in their words. It then sits in the notes, and tomorrow's brief reads it and stays quiet. Do not keep a separate list of things not to mention.

Non-negotiables

Only when config/season.md question 4 names them, and only when the floor has actually been slipping. One line, and only one.

The wrap writes a ## Non-negotiables block into each daily note. Read the last seven of them. Surface a line only on the morning a miss streak reaches exactly three nights running, or exactly seven. Name which one and how many nights, nothing else:

  • Asleep by 11 has missed 3 nights running

Those two numbers are what keep it from nagging. The brief has no memory of what it said yesterday, so "don't repeat yourself" is not a rule it can follow, but a four, five, or six night streak simply stays quiet. If they are holding, if fewer than four recent notes carry the block, or if the season names none, the section does not exist and you never bring it up. Do not report a floor that is holding and do not attach a suggestion. This is the user's own line, noticed at two points, not a habit tracker and not a nag.

Today's Schedule

Only when a calendar connector is available. Check for one before you write this section. If there is no calendar tool in the session, omit the section silently. Do not say the calendar is not connected, do not suggest connecting one, and never guess at the day's events from the notes.

When a calendar is live, pull today's events in the user's timezone and lay them out simply: time, what it is, who is in it. Flag anything that needs preparation the user has not done, for example a meeting with someone whose people/ note shows an open thread from last time. If the calendar returns nothing for today, omit the section. Silence reads better than "no meetings today."

Today's Priorities

Three items maximum. Pull from, in order:

  1. Anything already in Top of Mind Today.
  2. Tracker items due in the next two days that did not make that cut.
  3. What the user said about today in their most recent notes.
  4. What today's schedule implies, if you have a calendar.

If more than three feel urgent, the season breaks the tie. Write each as a thing to do, not a category. "Call the clinic back about the referral" beats "family admin."

Open Loops

Always a bullet list. Tracker items due in the next seven days that are not already above, plus unresolved threads from recent daily notes that never made it to the tracker. Order by urgency. Write each as a natural phrase rather than a task label.

Shape it like this:

  • Get back to the landlord about the lease renewal
  • Renew the studio membership before it lapses on Friday
  • Decide whether the Tuesday class is still worth the commute

If there are more than six, group the tail under one "also on the list" line.

Coming Up This Week

Two to four items, blended from tracker items due seven to fourteen days out, anything on the calendar later in the week that needs preparation now, and deadlines the user flagged in recent notes.

Only include what is genuinely significant or needs a head start today. Skip routine recurring events unless the user called one out.

Pattern Alert

Only when there is something real. A pattern needs at least three data points across recent days. Worth naming:

  • A feeling the user keeps reporting and never addresses.
  • Someone they keep saying they will contact and have not.
  • A habit they said mattered that has been missing for a week.
  • Energy or mood trending in one direction.

Check patterns/ first. If the theme already has a file, say it is continuing rather than presenting it as a discovery. If there is no real pattern, skip the section. Never manufacture one to have something to say.

This is an observation spoken out loud and nothing more. The briefing never writes to patterns/. Promotion is the weekly recap's call and it holds a much higher bar, so something worth mentioning here may never earn a file.

People on Your Radar

One to three people, only if someone came up recently in a way that suggests a follow up, a thank you, or a repair. One sentence each on why they surfaced. Use [[Name]] so the brief connects back to people/.

If an email connector is live, this is where a thread waiting on the user's reply belongs. One line, who and what. Without a connector, work from the notes.

One Thing to Remember

A single closing line. A quote from the user's own words, something they wrote that is worth hearing back, or a plain observation that fits this morning. It should sound like a friend who noticed something, not an affirmation generator.

Output

Two places: chat, then the Morning Brief page.

  1. Print the brief in chat, exactly as the rules below shape it.
  2. Publish it to the page. Read config/artifacts.md. pages: off: skip the page and say nothing. Otherwise check morning_brief_url.
    • Blank, in an interactive session: run "Setup" in artifacts/README.md for the Morning Brief first, then carry on.
    • Blank, unattended: skip the page and add Artifact: skipped (not set up yet) after the brief.
    • No Artifact or ArtifactData tool in the session: skip the page silently.
  3. Build the brief doc as "briefs/<YYYY-MM-DD>" in artifacts/README.md describes: the same sections you printed, in the same order, with the same words. Chat and page tell the same brief. Attach tracker_line to every item you took from the tracker, copied exactly (see "The tracker_line rule"). Add timezone, ask_context, and the calendar fields.
  4. Write it to tmp/brief.json with the Write tool, then ArtifactData set, collection: "briefs", doc_id = today's date, file_path = tmp/brief.json.
  5. On success, add one line after the brief: On the Morning Brief page. On failure: Artifact: failed (<reason>). Never retry in a loop and never let a failure here hold back the chat brief.

Do not write to daily/ and do not write to reviews/. The only commit is the tap write-back above, and a first-publish commit of config/artifacts.md. Do not send the brief anywhere else.

Anything the user tells you in reply is a new message, not part of the brief. A note or a "that's done" routes through the capture skill like any other, and capture does the writing.

Tone and format rules

Conversational, direct, warm enough to be readable. A smart friend who has read all your notes and has two minutes to give you the download. Not a dashboard, not a coach, not a productivity app.

Length target is two to three minutes read aloud. Err shorter.

For text to speech compatibility:

  • Write out abbreviations, "Monday" not "Mon"
  • Use digits for numbers, "3" not "three"
  • Prefer bullets, short sentences, and natural connective phrases
  • No special characters or symbols
  • No em dashes and no double hyphens anywhere, use commas, parentheses, or a new sentence
  • No spoken headers, transitions between sections should sound natural rather than announced
  • Avoid parenthetical asides that read badly out loud

For readability on screen:

  • Three to four sentences per paragraph, maximum
  • Whitespace between sections
  • Bold the section name where formatting exists, but write so each section still works when the formatting is stripped

What to avoid:

  • Filler openers. No "Great news", no "Here is your daily brief", no "Let's get into it"
  • Praise and encouragement. No "you're doing great"
  • Using the user's name more than once
  • Listing everything in the notes, curate aggressively
  • Any section that exists only to exist. Nothing real to say means skip it

Download now (ZIP)

Does it work?

Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.

No reviews yet

People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.

  • It worked
  • Partly
  • It didn't
What did you use it for? e.g. Reviewing pull requests in a Rails app

Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.

Your trail

Files you open will collect here, so you can find them again.

Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.

Now do this for your own.

Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.