cowork-inbox-triage — Claude Code skill: what it does, how to install
cowork-inbox-triage — a Claude Code skill that triages unread email via Microsoft 365 or Gmail MCP connectors. Scans threads from the last 24 hours and prepares drafts for review.
MIT 106 installs ★ 440 updated 2026-10-20
npx skills add OneWave-AI/claude-skills --skill cowork-inbox-triage -a claude-codeWhat cowork-inbox-triage does
Processes unread email threads and assigns each message to one of five categories, separating routine correspondence from noise and action items. When an email requires a decision, the skill drafts a recommended answer alongside the original quote. For standard replies, it writes drafts directly in the inbox but never hits send.
On the initial run, the skill examines 10–15 sent emails to match greeting habits, sentence length, and tone. It also tracks outbound conversations and flags threads where no reply arrived after three business days.
Requires a working MCP connector for Microsoft 365 or Gmail. If the email connector lacks write permissions, the skill generates pasteable text blocks inside the chat instead of native drafts.
What it helps with
- triage unread email threads from the past 24 hours
- draft replies in the sender's writing style without sending
- isolate messages that require human decisions and suggest answers
- track pending replies and draft follow-ups after three business days
Example prompts for the agent
- Triage my inbox
- Catch me up without creating drafts
- Who am I waiting on past the response window?
How to install cowork-inbox-triage
1Skills CLI (Vercel)
npx skills add OneWave-AI/claude-skills --skill cowork-inbox-triage -a claude-code2Manually: put the files in place
~/.claude/skills/cowork-inbox-triage/SKILL.mdCopy the whole skill folder; for one project use .claude/skills/cowork-inbox-triage/
1Skills CLI (Vercel)
npx skills add OneWave-AI/claude-skills --skill cowork-inbox-triage -a codex2Manually: put the files in place
.agents/skills/cowork-inbox-triage/SKILL.mdFor all projects use ~/.agents/skills/
1Skills CLI (Vercel)
npx skills add OneWave-AI/claude-skills --skill cowork-inbox-triage -a cursor2Manually: put the files in place
.cursor/skills/cowork-inbox-triage/SKILL.mdCursor also reads .agents/skills/ and .claude/skills/
1GitHub CLI
gh skill install OneWave-AI/claude-skills cowork-inbox-triage2Manually: put the files in place
.github/skills/cowork-inbox-triage/SKILL.md1Skills CLI (Vercel)
npx skills add OneWave-AI/claude-skills --skill cowork-inbox-triage -a gemini-cli2Platform CLI
gemini skills install https://github.com/OneWave-AI/claude-skills --path cowork-inbox-triageManually: put the files in place
.devin/skills/cowork-inbox-triage/SKILL.mdThe legacy .windsurf/skills/ path still works
Skills CLI (Vercel)
npx skills add OneWave-AI/claude-skills --skill cowork-inbox-triage -a opencodeClaude.ai
Zip the folder and upload it in Settings → Capabilities → Skills
Step by step, with commands for each platform: How to install skills in Claude Code: five ways
Questions about cowork-inbox-triage
How to install cowork-inbox-triage in Claude Code?
Run npx skills add OneWave-AI/claude-skills --skill cowork-inbox-triage -a claude-code in your terminal.
Does cowork-inbox-triage send emails automatically?
No. The skill strictly creates drafts or prints reply snippets in chat. The human owner sends every email.
What is required to run cowork-inbox-triage?
It needs an active Microsoft 365 or Gmail MCP connector with read and draft write access.
Source text
SKILL.md · MIT
--- name: cowork-inbox-triage description: Daily inbox triage using Cowork's email write tools (Microsoft 365 or Gmail) -- classifies unread mail, drafts replies in your voice, flags what needs a decision, and schedules follow-ups. Built to run as a recurring Cowork scheduled task. tools: Read, Write, Bash, WebSearch model: inherit --- # Cowork Inbox Triage Work an inbox the way a good chief of staff does: everything gets a disposition, nothing gets sent without approval, and the owner ends the run with a short list of real decisions instead of 80 unread threads. Uses the connected mail tools (Microsoft 365 or Gmail MCP connectors) for reading, labeling, and drafting. ## Setup On first run, learn the owner's voice: read 10-15 sent replies and note greeting style, sign-off, sentence length, and formality. Store the profile in the conversation and reuse it for every draft. Ask once which categories matter (default set below) and what the auto-archive rules are. ## Workflow 1. **Sweep.** Pull unread and flagged threads from the last cycle (default: 24 hours). Group by thread, not message. 2. **Classify** every thread into exactly one bucket: - `NEEDS DECISION` -- only the owner can answer (pricing, commitments, personnel) - `DRAFT READY` -- routine reply Claude can write for approval - `WAITING ON THEM` -- owner already replied; schedule a follow-up check - `FYI` -- no response needed; summarize in the digest - `NOISE` -- newsletters, notifications, cold outreach; label/archive per the standing rules 3. **Draft.** For every `DRAFT READY` thread, write the reply in the owner's voice as a draft -- never send. Keep drafts shorter than the email being answered. 4. **Follow-ups.** For `WAITING ON THEM` threads past their expected response window (default 3 business days), draft a polite bump. For threads with dates in them, propose calendar holds via the calendar write tools. 5. **Digest.** Post one summary: decisions needed (with one-line context and a recommended answer each), drafts awaiting approval, follow-ups queued, and a one-line count of what was archived. ## Rules - Never send email. Create drafts only; the human sends. - Never archive anything that mentions money, contracts, legal matters, or a specific person's problem -- misclassified noise is recoverable, a missed client email is not. - A recommended answer accompanies every `NEEDS DECISION` item. Surfacing a question without a recommendation is half the job. - Quote the sender's actual ask in the digest -- do not paraphrase requests into vagueness. - Respect confidentiality: HR, health, and personal matters get flagged privately, never summarized in a shared channel. - If the connector lacks write scope, degrade gracefully: produce the digest and paste-ready reply texts instead of drafts. ## Scheduled Mode As a Cowork scheduled task (weekday mornings), run the full workflow unattended in a cloud session and deliver the digest before the workday starts. Track threads across runs so a bump is never drafted twice for the same silence. ## Quick Commands - "Triage my inbox" -- full run, last 24h - "Catch me up" -- sweep and digest only, no drafts - "Draft replies to everything easy" -- classification + drafts only - "Who am I waiting on?" -- the `WAITING ON THEM` ledger with days elapsed
© OneWave-AI · MIT · OneWave-AI/claude-skills