agentleFS
serhii12serhii12/vault-os/.claude/skills/today/SKILL.md

today

Skill2 starsChanged 6 months agoMITView on GitHub

Start the day with a structured review and plan. Pulls calendar, email, tasks, recent messages, daily notes, and vault context, then synthesizes a core theme, top priorities, quick wins, and meeting audit, and finally generates an interactive HTML dashboard. Use when the user says "start my day", "today", "/today", "daily review", "morning review", "plan the day", or asks for a daily dashboard.

18 other projects use a version of this.

Adopted by

Most recently taken up by

Used with

Projects with it also use update and start.

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.
  • AnthropicSkill

    update

    Sync tasks and refresh memory from your current activity. Use when pulling new assignments from your project tracker into TASKS.md, triaging stale or overdue tasks, filling memory gaps for unknown people or projects, or running a comprehensive scan to catch todos buried in chat and email.

    Also in 3 projects that use this

  • AnthropicSkill

    start

    Set up your bio-research environment and explore available tools. Use when first getting oriented with the plugin, checking which literature, drug-discovery, or visualization MCP servers are connected, or surveying available analysis skills before starting a new project.

    Also in 3 projects that use this

The file

Show its rules and text

We split this skill into the rules it gives an agent

16 rules. 0 are common, 10 are distinctive, and 6 are very specific to this project.

  • 0

    Common

    Most of the 15 edited versions we read keep these.

  • 10

    Distinctive

    This author's own choices. Worth a look.

  • 6

    Very specific

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

  • Common
  • Distinctive
  • Very specific
  1. Distinctive:

    Run the full morning review and produce an interactive HTML dashboard.

    line 8
  2. Distinctive:

    Do not use Linear here.

    line 31
  3. Distinctive:

    Google Tasks MCP: if loaded — primary.

    line 46
  4. Distinctive:

    Run steps 1, 1.5, and 3 in parallel where possible.

    line 56
  5. Very specific:

    Sparse data handling: If Calendar is empty today, lean on vault patterns + overdue tasks + context files.

    line 71
  6. Distinctive:

    Use whichever Slack search/read tool is available in the session (see Prerequisites for how to find it).

    line 75
  7. Distinctive:

    If no Slack tool is loaded, skip and note in the synthesis output.

    line 77
  8. Very specific:

    Read the past 7 daily notes from Journal/daily/:

    line 81
  9. Distinctive:

    Read each. Look for: recurring themes, unfinished threads, commitments made, energy patterns, "I need to…" / "TODO" / "follow up on…" …

    line 87
  10. Distinctive:

    Sparse-data handling: If fewer than 3 daily notes exist in the last 7 days, expand to 14 days.

    line 89
  11. Very specific:

    Read in parallel: _agent/context/IDENTITY.md, _agent/context/USER.md, _agent/context/SOUL.md.

    line 93
  12. Distinctive:

    Check whether today has unusual structure that overrides normal ranking:

    line 97
  13. Very specific:

    Use the vault to enrich and challenge the preliminary priorities.

    line 118
  14. Very specific:

    Ensure the output directory exists: mkdir -p _agent/outputs.

    line 164
  15. Very specific:

    Write to _agent/outputs/today.html (overwrite).

    line 167
  16. Distinctive:

    No hedging. If signal is weak, say so once.

    line 202

“Common” is measured: kept by at least 6 in 10 of the 15 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

Today — Daily Review & Planning

Run the full morning review and produce an interactive HTML dashboard. The user starts their day on this output, so it must be complete and decisive — not exploratory.

Conventions

  • Daily notes: Journal/daily/YYYY-MM-DD.md.
  • Context files: _agent/context/IDENTITY.md, _agent/context/USER.md, _agent/context/SOUL.md, _agent/context/AGENTS.md.
  • Dashboard output: _agent/outputs/today.html (overwrite each run; directory already exists).
  • Date: today, in the user's local TZ. Get the day name with date +%A.
  • Tag any file you create with source: agent in frontmatter.

Prerequisites (MCP servers)

The user has these connected:

  • Google Calendar (mcp__*Google_Calendar__*) — required.
  • Gmail (mcp__*Gmail__*) — required.
  • Slack — required for Step 1.5. The exact tool name varies by client. At session start, scan available tools for one matching mcp__*[Ss]lack* (e.g., search_messages, read_channel). If none is loaded, skip Step 1.5 and tell the user once to expose the Slack MCP — do not silently proceed without it.

These are optional / may not be present:

  • Google Tasks — preferred for the task system. If no Google Tasks MCP is loaded, fall back to Obsidian Tasks (see below), then to free-text TODO scanning. Tell the user once that Google Tasks would give richer data. Do not silently skip.
  • Granola (mcp__*Granola__*) — optional, for meeting notes from yesterday.

Do not use Linear here.

Task source priority

The vault has the Obsidian Tasks plugin installed. Tasks across markdown files use this syntax:

  • - [ ] Task text 📅 2026-05-06 — due date
  • - [ ] Task text ⏳ 2026-05-06 — scheduled
  • - [ ] Task text 🛫 2026-05-06 — start date
  • - [x] Done ✅ 2026-05-06 — completed
  • Priorities: ⏫ high, 🔼 med-high, 🔽 low, ⏬ lowest
  • Recurring: 🔁 every week

Order of precedence for Step 1 task discovery:

  1. Google Tasks MCP if loaded — primary.
  2. Obsidian Tasks markdown — grep for open - [ ] items with 📅 dates and bucket by date. Skip _agent/, 4-Archive/, .obsidian/, .claude/. Example:
    grep -rn "^- \[ \].*📅" --include="*.md" \
      1-Projects 2-Areas 3-Resources Inbox Journal Goals Content People Books Recipes 2>/dev/null
  3. Free-text TODOs in the past 7 daily notes — last resort.

Workflow

Run steps 1, 1.5, and 3 in parallel where possible. Steps 4 onward depend on prior context.

Step 1 — Calendar + Email + Tasks (parallel)

Run these concurrently:

  1. Google Calendar — list events for today + next 3 days.
  2. Gmail — search threads with newer_than:1d for unread/important threads. Look for: requests, commitments, time-sensitive items.
  3. Tasks — pull from the highest-priority source available (see "Task source priority" above) and bucket them:
    • Overdue (dueDate < today) — flag prominently
    • Due today (dueDate == today)
    • Due this week (today < dueDate ≤ today+7)
    • Quick wins — low-effort tasks (short titles; verbs like "send", "reply", "update", "confirm", "ping"; or tagged #quick / quick priority 🔽)
    • No due date (floating)

Sparse data handling: If Calendar is empty today, lean on vault patterns + overdue tasks + context files. If the task source returns few results, also scan the past 7 daily notes for informal task mentions (TODO, I need to…, follow up on…).

Step 1.5 — Recent messages (Slack, last 24h)

Use whichever Slack search/read tool is available in the session (see Prerequisites for how to find it). Query for the last 24h of DMs and channels the user is active in. Look for: commitments made, plans confirmed, requests from others, follow-ups needed.

If no Slack tool is loaded, skip and note in the synthesis output.

Step 2 — Review recent daily notes

Read the past 7 daily notes from Journal/daily/:

ls Journal/daily/ | sort -r | head -7

Read each. Look for: recurring themes, unfinished threads, commitments made, energy patterns, "I need to…" / "TODO" / "follow up on…" mentions.

Sparse-data handling: If fewer than 3 daily notes exist in the last 7 days, expand to 14 days. If still sparse, note it explicitly in the synthesis and rely more on calendar + context files.

Step 3 — Load context files

Read in parallel: _agent/context/IDENTITY.md, _agent/context/USER.md, _agent/context/SOUL.md. If the user's recent notes name an active project, also read its folder under 1-Projects/. Read _admin/CONNECTIONS.md if the day's priorities touch database/integration work.

Step 4 — Detect hard constraints

Check whether today has unusual structure that overrides normal ranking:

  • Solo parenting / partner travel changing childcare?
  • Travel day or transit compressing the work window?
  • Recording / content day requiring energy management?
  • Special events (launches, deadlines, investor meetings)?
  • Back-to-back meetings with no gaps?

Pull signals from calendar event titles/locations and any explicit notes in today's daily-note (if it exists). Flag constraints prominently — they override priority ranking in Step 7.

Step 5 — Preliminary priorities

Synthesize so far:

  • 2–3 most pressing items from calendar + tasks + recent notes
  • Dominant themes from the past week
  • Active tensions or open questions from context files

These guide vault exploration in Step 6 — don't lock them in yet.

Step 6 — Deep vault exploration

Use the vault to enrich and challenge the preliminary priorities. Use Grep and Read directly:

  1. Active themes — Count Obsidian inline #tag frequency across active folders and compare with today's preliminary priorities. Disconnect = signal.
    grep -rho "#[a-zA-Z][a-zA-Z0-9/_-]*" --include="*.md" \
      Journal/daily/ 1-Projects/ 2-Areas/ 2>/dev/null \
      | sort | uniq -c | sort -rn | head -15
  2. Priority-informed search — For each top priority, Grep across 1-Projects/, 2-Areas/, 3-Resources/, Journal/ for the topic. Read 2–3 most relevant matches.
  3. Backlinks (2–3 hops) — For key notes surfaced, follow [[wikilinks]] (including aliased [[note|alias]]) outward. Use grep -rl --include="*.md" "\[\[<note-name>\(\]\]\||\)" or simpler: grep -rl "\[\[<note-name>" --include="*.md".
  4. Orphans / forgotten notes — Quick scan of Inbox/ and recently modified notes outside the agent's own outputs:
    find Inbox 1-Projects 2-Areas 3-Resources Journal -name '*.md' -mtime -14 2>/dev/null
    Skip _agent/, 4-Archive/, .obsidian/, .claude/. Filter to notes with no source: frontmatter (i.e., human-authored).
  5. Surface contradictions — Does a recent daily note express doubt about a priority? Does a context-file open question suggest a different focus? Has thinking on a key topic shifted?

The goal is illumination, not exhaustive cross-referencing. Stop when you have signal.

Step 7 — Synthesize

Produce, in order:

Core Theme — single phrase or sentence capturing what today should be about.

Top 2 Priorities — the two most important things. Rank by:

  • Blocks others if not done
  • Has a hard deadline
  • Compounds negatively if delayed
  • Directly advances current goals from context files

For each, include the vault context: related thinking found, connections, questions raised.

Quick Wins (5) — tasks completable in <15 min that unblock someone, clear mental overhead, or prevent small things from becoming big.

Meeting Audit — flag meetings that should be:

  • Cancelled — no agenda, input not essential, could be async
  • Converted to async — replaceable by doc/video/thread
  • Shortened — padded time

For each flagged meeting, suggest the specific action and the artifact that would replace it.

Step 8 — Generate the dashboard

Generate _agent/outputs/today.html from the template at assets/dashboard-template.html. See references/dashboard-spec.md for placeholder names and data shape. Steps:

  1. Ensure the output directory exists: mkdir -p _agent/outputs.
  2. Read assets/dashboard-template.html.
  3. Replace placeholders with synthesized data (date, theme, priorities, schedule, tasks, upcoming events, alerts).
  4. Write to _agent/outputs/today.html (overwrite).
  5. Print the absolute file path so the user can open it.

The dashboard is the deliverable. The text synthesis precedes it; the HTML is what the user lives in.

Output to chat (before the dashboard)

Format with clear headers in this order:

## Core Theme
<one phrase>

## Hard Constraints
<bulleted, only if present>

## Top 2 Priorities
1. <priority> — <vault context>
2. <priority> — <vault context>

## Quick Wins
- ...

## Meeting Audit
- <meeting> → cancel / async / shorten — <reason>

## Dashboard
Generated at: _agent/outputs/today.html

Then mention any sparse-data caveats from Step 2.

Style

  • Direct, actionable, terse. Match Noa's voice (see _agent/context/SOUL.md).
  • No hedging. If signal is weak, say so once; don't over-qualify every line.
  • Action-phrased bullets, no trailing punctuation.

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.