---
name: directory_submissions
description: Run product directory submissions end to end as a launch channel: pick the right startup, SaaS, AI, and agent directories, build a paste-ready submission kit, track every submission through review, and follow up until it is live. Use when launching a product, tool, agent, or side project through directories, or when planning a launch-week push.
---

# Directory Submissions

## Purpose

Turn "submit to some directories" into a managed launch channel. Most directory submissions die in one of three places: the wrong directories (indie launch board for a B2B tool, AI catalog for a non-AI product), copy pasted from one form to the next until editors reject it, or no follow-up after submitting so the listing never goes live. This skill prevents all three with a lane-based target list, a copy kit written once and tuned per directory, and a tracker that tells you exactly what to chase each week.

## Workflow

1. **Classify the product.** Pick which of the five lanes apply (see `references/directory-catalog.md`):
   - Launch boards (Product Hunt and indie launchers) if there is a real launch moment
   - Startup directories (BetaList, Crunchbase, launch kits) if the product is new or pre-launch
   - SaaS review directories (G2, Capterra, SaaSHub, AlternativeTo) if it is sellable software with users
   - AI directories (There's An AI For That, Futurepedia, Toolify) only if the product genuinely uses AI
   - Agent directories if it is an agent, agent tool, or MCP server
   Skip lanes that do not fit. Ten well-matched listings beat fifty sprayed ones.

2. **Triage the targets.** For each chosen lane, pull the directory catalog and mark: free vs paid tier, current price (re-check the directory's own submit page; prices move), review time, and any conditions (badge, backlink, account). Budget 5 to 10 directories per week so follow-up stays manageable.

3. **Build the copy kit.** Write one master asset pack (see `references/copy-kit.md`): name, URL, tagline, short description, long description, category, tags, pricing, logo, screenshots, demo video. Then tune the description per directory. Editors notice recycled copy, and several directories penalize identical text across listings.

4. **Submit in priority order.** High-fit free listings first, then paid tiers you have explicitly approved. Screenshot or save every confirmation; a surprising number of directories show no clear confirmation screen.

5. **Track and follow up.** Log every submission with `bin/dirsub.py`. Each week, run the follow-up check: anything submitted longer than the directory's stated review time with no confirmation gets a follow-up email or resubmission. A submitted-not-live listing is not a listing.

## Output Contract

Every run produces:
- A ranked target list (directories with lane, cost, review time, submit URL)
- A copy kit (master assets + per-directory tuned descriptions)
- A tracking sheet (`bin/dirsub.py` records, one row per submission with status, dates, follow-up due)
- A weekly follow-up list until every target is live or explicitly dropped

## Operating Rules

1. Never pay a directory without the user's explicit word on that exact tier. Free tiers go ahead; anything paid stops and asks.
2. Never fabricate numbers (users, revenue, pricing, ratings) in listing copy. Unverifiable claims get cut.
3. Never submit a product to a lane it does not fit just to fill the sheet. Off-lane submissions waste the reviewer's time and yours.
4. Re-verify prices and submit URLs on the directory's own page before filling a form. Directories change tiers without notice.
5. One listing per product per directory. Duplicates get removed and can burn the account.
6. Log the confirmation for every submission. If the last click does not prove it went through, check back before submitting again.
7. Follow-ups are the job, not a bonus step. A tracker full of "submitted, pending" with nothing live is a failed run.
8. If a directory requires a badge or backlink on the user's site, flag it before submitting; it is the user's call whether to accept that condition.