How do I install the Inbox Cleaner skill?

Run npx skillmds@latest add mhassan0000/inbox-cleaner in your terminal (requires Node.js), paste this page's agent-chat prompt into Claude, Cursor, or any MCP-connected agent, or download the SKILL.md file and copy it into your agent's skills directory.

Is Inbox Cleaner safe to use?

This skill has not completed SkillMD's automated safety review yet. Independent scanners report: SkillSpector: PASS, Skill Scanner: PASS. Capability flags: makes network calls, reads secrets. SkillMD never runs a skill's scripts for you; review the SKILL.md before installing.

Which AI agents work with Inbox Cleaner?

This skill is tagged as working with Claude Code, Claude.ai, OpenAI Codex. SKILL.md is an open format, so most agents that read a skills directory can load it too.

Is Inbox Cleaner free to use?

Yes. Installing skills from SkillMD is free, and the skill stays under its author's original license.

Who published Inbox Cleaner?

MHassan0000 (@mhassan0000) published this skill. Their other Agent Skills are listed on their SkillMD profile.

Inbox Cleaner by mhassan0000 — Agent Skill

Name: Inbox Cleaner
Rating: 4.2 (5 reviews)
Author: MHassan0000

Inbox Cleaner

Clean up Gmail inbox by reading all unread emails, using AI to identify which ones are genuinely important (personalized, human-written), and marking the rest as read. Use when cleaning inbox, triaging email, or clearing unread notifications.

MHassan0000 Updated Sep 8, 2026 1 repo stars

File contents

Inbox Cleaner — AI-Powered Unread Email Triage

Goal

Go through every unread email in nick@leftclick.ai, classify each one as important or not, and mark unimportant ones as read. Only truly personalized, human-written emails survive as unread.

What Counts as Important

An email is important ONLY if it was clearly written by a human specifically for Nick. Indicators:

References specific details about Nick, his business, or prior conversations
Contains substantive content that couldn't be a template
Comes from a known contact with a real message
Replies in an existing thread with genuine human input

What Gets Marked as Read (Not Important)

Automated notifications (GitHub, Stripe, Slack digests, calendar, etc.)
Marketing emails and newsletters
Cold outreach / sales emails (templated pitches, "I noticed your company..." spam)
Service alerts, receipts, shipping updates
Social media notifications
Automated replies (out-of-office, delivery confirmations)
Mass emails from SaaS products
Any email that feels like it was sent to 100+ people

Process

Step 1: Fetch Unread Emails

python3 execution/inbox_cleaner.py --fetch

Fetches all unread emails from nick@leftclick.ai and saves metadata + snippets to data/inbox_unread.json.

Step 2: Classify Emails

python3 execution/inbox_cleaner.py --classify

Uses Claude to classify each email as "important" or "not_important". Saves results to data/inbox_classified.json.

Step 3: Review Classification

python3 execution/inbox_cleaner.py --review

Displays the classification results for review before marking as read.

Step 4: Mark as Read

python3 execution/inbox_cleaner.py --mark-read

Marks all "not_important" emails as read. Important emails remain unread.

Gmail Auth

Token: config/token_leftclick.json
Credentials: config/credentials.json
Scopes: gmail.modify (read + mark as read)
Email: nick@leftclick.ai

Output

data/inbox_unread.json — fetched unread emails
data/inbox_classified.json — classification results with reasoning
Console summary showing what was kept vs marked read

Edge Cases

If classification is uncertain, keep the email as unread (err on the side of caution)
Emails from contacts in existing threads should be treated as important even if short
Forwarded emails: judge based on the forwarding context, not the forwarded content

First-Run Setup

Before executing, check if the workspace has a .gitignore file. If it doesn't, assume the user is new to this skill. In that case:

Ask the user if this is their first time running this skill
If yes, walk them through how it works and what they need to configure/set up (API keys, env vars, dependencies, etc.)
Let them know that Nick wishes them the best!

---name: inbox-cleanerdescription: Clean up Gmail inbox by reading all unread emails, using AI to identify which ones are genuinely important (personalized, human-written), and marking the rest as read. Use when cleaning inbox, triaging email, or clearing unread notifications.---# Inbox Cleaner — AI-Powered Unread Email Triage## GoalGo through every unread email in nick@leftclick.ai, classify each one as important or not, and mark unimportant ones as read. Only truly personalized, human-written emails survive as unread.## What Counts as ImportantAn email is important ONLY if it was clearly written by a human specifically for Nick. Indicators:- References specific details about Nick, his business, or prior conversations- Contains substantive content that couldn't be a template- Comes from a known contact with a real message- Replies in an existing thread with genuine human input## What Gets Marked as Read (Not Important)- Automated notifications (GitHub, Stripe, Slack digests, calendar, etc.)- Marketing emails and newsletters- Cold outreach / sales emails (templated pitches, "I noticed your company..." spam)- Service alerts, receipts, shipping updates- Social media notifications- Automated replies (out-of-office, delivery confirmations)- Mass emails from SaaS products- Any email that feels like it was sent to 100+ people## Process### Step 1: Fetch Unread Emails```bashpython3 execution/inbox_cleaner.py --fetch```Fetches all unread emails from nick@leftclick.ai and saves metadata + snippets to `data/inbox_unread.json`.### Step 2: Classify Emails```bashpython3 execution/inbox_cleaner.py --classify```Uses Claude to classify each email as "important" or "not_important". Saves results to `data/inbox_classified.json`.### Step 3: Review Classification```bashpython3 execution/inbox_cleaner.py --review```Displays the classification results for review before marking as read.### Step 4: Mark as Read```bashpython3 execution/inbox_cleaner.py --mark-read```Marks all "not_important" emails as read. Important emails remain unread.## Gmail Auth- Token: `config/token_leftclick.json`- Credentials: `config/credentials.json`- Scopes: `gmail.modify` (read + mark as read)- Email: nick@leftclick.ai## Output- `data/inbox_unread.json` — fetched unread emails- `data/inbox_classified.json` — classification results with reasoning- Console summary showing what was kept vs marked read## Edge Cases- If classification is uncertain, keep the email as unread (err on the side of caution)- Emails from contacts in existing threads should be treated as important even if short- Forwarded emails: judge based on the forwarding context, not the forwarded content## First-Run SetupBefore executing, check if the workspace has a `.gitignore` file. If it doesn't, assume the user is new to this skill. In that case:1. Ask the user if this is their first time running this skill2. If yes, walk them through how it works and what they need to configure/set up (API keys, env vars, dependencies, etc.)3. Let them know that Nick wishes them the best!

Frequently asked questions