mail-action-items
Skill7 starsChanged 9 months agoMITView on GitHub
Extract action items, tasks, and to-dos from recent emails. Scan email bodies for requests, deadlines, approvals needed, and follow-ups. Use when user wants to know what they need to do based on their email, or asks "what do I need to act on?" Arguments: optional time range (default: last 3 days) or account filter.
In this repository32 other files it gives its agents.
- calendar-accountsSkill
- calendar-attendeesSkill
- calendar-conflictsSkill
- calendar-coreSkill
- calendar-eventSkill
- calendar-freeSkill
- calendar-recurringSkill
- calendar-searchSkill
- calendar-statsSkill
- calendar-todaySkill
- calendar-upcomingSkill
- mail-accountsSkill
- mail-attachmentsSkill
- mail-bankingSkill
- mail-contactsSkill
- mail-coreSkill
- mail-digestSkill
- mail-expensesSkill
- mail-fromSkill
- mail-meetingsSkill
- mail-needs-replySkill
- mail-newsletterSkill
- mail-readSkill
- mail-searchSkill
- mail-securitySkill
- mail-statsSkill
- mail-subscriptionsSkill
- mail-threadSkill
- mail-top-sendersSkill
- mail-travelSkill
- mail-triageSkill
- mail-workSkill
The file
We split this skill into the rules it gives an agent
1 rule. About 0 are common, 1 is distinctive, and 0 are very specific to this project.
≈0
Common · estimate
Many files like this have a rule of this kind.
≈1
Distinctive · estimate
This author's own choices. Worth a look.
0
Very specific
Tied to their project's names, files or tools. Probably skip.
- Common
- Distinctive
- Very specific
- Distinctive: line 80
Look for action-item patterns in the body text:
≈ “Common” is an estimate: at least 1 in 5 of the 350,495 skill files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.
Files
Mail Action Items — Extract Tasks from Email
Time window: $ARGUMENTS (default: last 3 days)
Step 1: Find candidate emails with action signals
Subject-line signals
DB="$HOME/Library/Mail/V10/MailData/Envelope Index"
SINCE=$(($(date +%s) - 259200)) # 3 days default; adjust per $ARGUMENTS
sqlite3 "$DB" "
SELECT datetime(m.date_received,'unixepoch','localtime') as dt,
s.subject, a.address as sender, a.comment as name,
mb.url as mailbox, m.ROWID, m.read, m.is_urgent
FROM messages m
JOIN subjects s ON m.subject = s.ROWID
JOIN addresses a ON m.sender = a.ROWID
JOIN mailboxes mb ON m.mailbox = mb.ROWID
WHERE m.date_received >= ${SINCE}
AND m.deleted = 0
AND m.automated_conversation != 2
AND mb.url NOT LIKE '%Spam%' AND mb.url NOT LIKE '%Trash%'
AND mb.url NOT LIKE '%Sent%'
AND (
s.subject LIKE '%action required%'
OR s.subject LIKE '%please review%'
OR s.subject LIKE '%approval%'
OR s.subject LIKE '%approve%'
OR s.subject LIKE '%sign off%'
OR s.subject LIKE '%deadline%'
OR s.subject LIKE '%due%'
OR s.subject LIKE '%RSVP%'
OR s.subject LIKE '%respond by%'
OR s.subject LIKE '%feedback%'
OR s.subject LIKE '%request%'
OR m.is_urgent = 1
OR m.flagged = 1
)
ORDER BY m.is_urgent DESC, m.date_received DESC;" 2>/dev/nullUrgent/flagged unread (always surface)
sqlite3 "$DB" "
SELECT datetime(m.date_received,'unixepoch','localtime') as dt,
s.subject, a.address as sender, m.ROWID
FROM messages m
JOIN subjects s ON m.subject = s.ROWID
JOIN addresses a ON m.sender = a.ROWID
JOIN mailboxes mb ON m.mailbox = mb.ROWID
WHERE m.date_received >= ${SINCE}
AND m.read = 0 AND m.deleted = 0
AND m.automated_conversation != 2
AND m.unsubscribe_type = 0
AND mb.url NOT LIKE '%Spam%' AND mb.url NOT LIKE '%Trash%'
AND mb.url NOT LIKE '%Sent%'
ORDER BY m.date_received DESC;" 2>/dev/nullStep 2: Read bodies and extract action items
For each candidate email, parse the body:
python3 ~/.claude/skills/_mail-shared/parser.py <ROWID1> <ROWID2> ...Look for action-item patterns in the body text:
- "Please [verb]..." → direct request
- "Could you..." / "Can you..." → request
- "By [date]" / "Before [date]" / "Due [date]" → deadline
- "Let me know..." → response expected
- "Your approval is needed" / "Please approve" → approval required
- "Next steps:" / "Action items:" / "TODO:" → explicit task lists
- "Waiting on you" / "Pending your" → you're the blocker
- Meeting notes with "Suggested next steps" or your name + "will"
Step 3: Deduplicate and prioritize
- Group by sender/thread
- Prioritize: overdue deadlines > urgent flagged > approval requests > general requests
- For Jira/project management alerts, extract ticket ID and description
Output Format
Action Items from Your Email (last 3 days)
For each item:
- Task description — from [Sender], received [date]
- Context: [1-line summary from email body]
- Deadline: [if mentioned]
Group by: Overdue → Today → This week → No deadline Mark items from unread emails with [UNREAD].
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
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.

