inbox-cleanup

Inbox Cleanup Skill

A playbook for large-scale email inbox cleanup. The core insight: sender-based scans are low-recall. Subject/body pattern queries catch 25x more archivable email. This skill is a multi-pass pipeline built around that insight.

Works with any connected email provider. Adapt query syntax to whatever the provider supports — the strategy (what to search for, how to decide what to archive) is universal.

Gmail is a required integration. It's declared via includes: ["gmail"] in the frontmatter so it loads synchronously on activation, not lazily after the preferences form. Load/confirm the Gmail integration the moment this skill activates — before Phase 1 — so a missing or unauthorized connection surfaces up front rather than mid-cleanup.

---

Phase 1: Preference Capture

Do this before touching anything. Ask the user:

1. Aggressiveness level

• Conservative — newsletters with unsubscribe headers + obvious spam only
• Standard — above + cold outreach heuristics (subject patterns, unknown senders)
• Aggressive — above + anything from senders with no prior thread history

2. Age threshold Archive everything older than X days? Common choices: 30 / 60 / 90 days. Or no age filter.

First-run scope: On first invocation, scope to last 30 days or top 3 noise patterns, whichever surfaces faster. Show result, offer to expand. Prove the approach on a fast, visible slice before draining the whole backlog.

3. VIP senders to protect Ask: "Are there any senders that might look like cold outreach but you actually care about? Think: specific individuals at investors, advisors, your lawyer, accountant, recruiters you're actively working with."

Build an explicit keep list. Do not archive anything matching it, ever, regardless of aggressiveness.

4. Categories to confirm before archiving These need a sample + explicit approval before bulk action:

• Financial/billing alerts
• Legal/contracts
• Account suspension notices
• Government/regulatory

---

Phase 2: Urgency Triage (do this before any archiving)

Scan the inbox first for high-stakes items that should be surfaced, not archived. Look for:

Surface these to the user before running the cleanup. They're easy to miss buried in a big inbox.

---

Phase 3: The Query Pipeline

Run these passes in order. Each pass should paginate to exhaustion (keep fetching while more results exist). After each pass, show the user a count + 5 sample subjects before archiving anything.

Pass 1: Age-based bulk (biggest catch)

Search for all inbox messages older than the user's age threshold (e.g. 30 days). Typically 50–80% of the archivable backlog. Always show a sample before bulk archiving.

Note on result caps: Some providers cap query results (e.g. ~5,000). If a query returns exactly at the cap, archive that batch and re-run the same query — the next batch will surface. Repeat until it returns fewer than the cap.

Pass 2: Personalized cold outreach (subject patterns)

Ask the user for their first name and company name, then search for subject lines containing patterns like:

• [FirstName] -, [FirstName],, for [FirstName], hi [FirstName], hey [FirstName], [FirstName] |
• [CompanyName] -, [CompanyName]?, for [CompanyName], re: [CompanyName], [CompanyName] AI

These are the highest-recall patterns for cold outreach and partnership spam. A startup founder's inbox will see the biggest wins here.

Pass 3: Generic cold outreach phrases

Search for subject lines containing:

• "quick question", "quick note", "checking in"
• "following up", "just following up", "circling back"
• "would love to connect", "15 minutes", "quick call"
• "intro call", "reaching out", "came across your"
• "are you the right person", "happy to hop on"

Pass 4: No-reply & newsletters

Search for:

• Messages from noreply/no-reply/donotreply sender addresses
• Subject lines containing "unsubscribe", "newsletter", "weekly digest", "monthly digest"

Pass 5: Calendar noise

Search for subject lines containing:

• "accepted:", "declined:", "tentative:"
• "has accepted", "has declined", "invitation:"

Calendar response confirmations are pure noise. Safe to bulk archive without review.

Pass 6: Transactional/receipts

Search for subject lines containing:

• "your order", "order confirmation", "your receipt"
• "shipment", "has shipped", "delivered"

Cross-check against urgency triage first — filter out any "past due" or "final notice" items before archiving this batch.

Pass 7: Sketchy TLDs

Search for messages from sender domains ending in .shop, .biz, .xyz, .info, .club, .online.

Disproportionately spam. Safe to bulk archive.

Pass 8: High-volume repeat senders

After the above passes, run a sender frequency count on what remains. Any sender with 3+ emails not on the keep list is a candidate for bulk archive. Show grouped list to user for approval.

---

Phase 4: Cold Outreach Classification

For emails not caught by pattern queries, use LLM-based classification in Standard/Aggressive mode. Flag as cold outreach if 3+ signals are present:

• Sender domain not in user's contact/thread history
• No prior reply from user to this sender
• Subject contains user's name + company together (personalization ≠ trust)
• Body contains: "came across your company", "I help companies like", "reaching out because", "15 minutes", "quick call", "are you the right person"
• Sender domain is a known outreach tool: apollo.io, outreach.io, lemlist.com, instantly.ai, salesloft.com
• Email is not a reply (no Re: prefix, no quoted text from user in body)

---

Dry-Run Defaults

Every bulk archive previews before it executes — regardless of batch size or trust stage. Run the pipeline with --dry-run on all archive calls, then render a ui_show table preview the user commits or refines from. Never archive in bulk straight from a query.

The preview table must show:

1. Total emails matching — the full count this bulk archive would touch
2. Top-10 sender breakdown — senders by volume, so the user spots anything they care about
3. 10–20 sample subjects — a representative spread, not just the first few
4. Categories flagged for confirm-before-archive — the Phase 1 categories (financial/billing, legal/contracts, account suspension, government/regulatory) that matched, called out for explicit approval

Surface "things worth flagging before you confirm" inside the preview, not after. If the dry-run catches claim documents, failed-payment notices, or any urgency-triage signal (Phase 2), call them out in the preview so the user sees them while deciding — never let a flag-worthy item get archived first and surfaced afterward.

After rendering the preview:

1. Ask the user to confirm or refine: "This would archive X,XXX emails across Y passes. Commit, or refine the scope?"
2. If confirmed, commit via bun run scripts/gmail-commit.ts commit --run-id "<run-id>"
3. If rejected, cancel via bun run scripts/gmail-commit.ts cancel --run-id "<run-id>"

Larger batches (e.g. >1,000 operations) and lower trust stages (stage 0 flag-only) warrant extra scrutiny in the preview, but the preview itself is always required before any bulk archive — including small batches and high trust stages. Direct archives are still logged for audit/reversal.

---

Error Recovery & Resume

Archive operations are logged to an operation log for resumability. If a pass fails mid-run (rate limit, daily quota, OAuth expiry, crash):

1. Check for interrupted runs before starting a new cleanup: bun run scripts/gmail-runs.ts list. If a recent run shows status: "interrupted", offer to resume it.
2. Resume: bun run scripts/gmail-archive.ts archive --resume "<run-id>". This skips already-committed chunks and retries pending ones.
3. Daily quota (403): The archive script detects daily quota exhaustion and writes an interrupted log entry with a resume hint. Do not retry until after midnight PT — offer to resume the run later.
4. Rate limit (429): Handled automatically with exponential backoff (up to 5 retries for batch operations). No user intervention needed.

All archive outputs now include a run_id. Pass --run-id to group multiple passes under one run, and --phase to label the pipeline phase (e.g. --phase "noise_archive").

---

Phase 5: Post-Cleanup

1. Report totals — how many archived per pass, which categories, and the run_id for each pass
2. Update blocklist — remember which senders/domains were archived; use for faster future passes
3. Surface any urgents found — if financial/legal/suspension items surfaced during the pass, present them now with recommended actions
4. Mention reversal — remind the user: "If any of these archives were wrong, I can reverse specific threads: bun run scripts/gmail-reverse.ts --run-id <id> --thread <message-id>"

---

Phase 6: Permanent Filter Setup

After cleanup, propose Gmail filters so the same categories don't re-accumulate. This bridges cleanup (drain backlog once) and inbox-management (keep inbox clean on schedule).

Note: Filter creation capabilities vary by provider. The gmail-auto-filters.ts script handles Gmail. If the provider doesn't support programmatic filter creation, give the user manual instructions instead.

Filters are permanent behavior changes. Unlike a one-time archive, a filter silently skips the inbox for every future matching email. A wrong filter means the user misses emails they were expecting — with no indication anything happened. Always confirm with the user before creating filters.

Which patterns are safe as permanent filters

One-time bulk archiving and permanent auto-archiving are different risk levels. The auto-filter script only derives candidates from patterns marked "Yes" below:

Running auto-filter generation

After the cleanup pipeline completes (Phase 5 post-cleanup report), invoke:

# Preview: show what filters would be created (no confirmation prompt)
bun run scripts/gmail-auto-filters.ts preview --run-id "<cleanup-run-id>"

# Generate: show plan, confirm with user, then create
bun run scripts/gmail-auto-filters.ts generate --run-id "<cleanup-run-id>"

If --run-id is omitted, the script finds the most recent completed cleanup run automatically.

The script:

1. Reads the cleanup run's op-log to extract archived patterns
2. Derives filter candidates from safe categories only
3. Fetches existing Gmail filters and skips duplicates
4. Shows the user a confirmation dialog listing every filter that will be created, its criteria, and its label — the user must explicitly approve before any filter is created
5. Creates one filter per logical category with an auto/* label (e.g. auto/no-reply, auto/calendar, auto/newsletter, auto/sketchy-tld)
6. Logs all filter creations to the op-log for audit and reversal

Label strategy

Every auto-filter applies an auto/* label instead of silently archiving. This gives the user an audit trail — search label:auto/calendar to see what was caught. Labels are created automatically if they don't exist.

After filter creation

Tell the user:

• How many filters were created and what each covers
• How to find auto-archived emails (search by label, e.g. label:auto/no-reply)
• How to remove a filter: bun run scripts/gmail-manage.ts filters --action delete --filter-id "<id>"

---

Reference: Proven Catch Rates

From a single cleanup session on a startup founder's inbox (April 2026):

Total: ~90,000+ emails in one session. The name/company pattern passes alone accounted for ~85k. This is why patterns dominate sender scans.