REACH DOG FREE CATEGORY AGENT
Tee Catalog Connection Agent
Source: https://reach.dog/selling-guides/how-to-sell-womens-t-shirts-online
Category snapshot 2026-08-10 | Prompt version 2026-09-08

HOW TO USE
1. Copy these instructions into the LLM you choose.
2. Replace [PASTE ONE PRODUCT RECORD] with one current, verified product record.
3. Review every relevance decision, fact trace, and draft before publishing.

The prompt blueprint is free. Model access and usage depend on the provider you choose. This prompt does not publish, schedule, or change a catalog.

CATEGORY
women's t-shirts

SIX BUYER QUESTIONS
1. What makes a women's t-shirt useful for everyday layering?
   Required facts: fabric; fit; care; size
   Boundary: Do not call every women's t-shirt a layering base because the category includes the signal.
2. Which women's t-shirts suit casual outings in summer?
   Required facts: sleeve; color; measurements; construction
   Boundary: Casual outings and summer are category signals; they do not establish cooling, comfort, or weather performance.
3. When should I feature women's t-shirts for spring and summer?
   Required facts: season-relevant construction; fabric content; sleeve length; care instructions
   Boundary: Observed season labels do not prove that every product is suitable for a particular climate or season.
4. How can I choose a women's t-shirt for a young adult or casual wearer?
   Required facts: size range; fit; measurements; garment identity
   Boundary: Audience co-occurrence does not establish age-specific fit, preference, or suitability for every wearer.
5. What facts support saying a women's t-shirt is comfortable?
   Required facts: fiber content; size range; care instructions; benefit evidence
   Boundary: Comfort, softness, and breathability in category metadata are prompts, not universal claims about women's t-shirts.
6. How do I know this listing is actually for a women's t-shirt rather than a similar garment?
   Required facts: garment identity; title; neckline; sleeve length
   Boundary: A similar title or category match is not proof of garment identity; the sampled identity issue is bounded and not a category-wide error rate.

PRODUCT RECORD
[PASTE ONE PRODUCT RECORD]

PRODUCT RECORD TEMPLATE
Write unknown for missing or undocumented facts.
Product title:
Product description as published:
Facts this category needs: fabric; fit; care; size; sleeve; color; measurements; construction; season-relevant construction; fabric content
Care or usage guidance:
Documentation on hand:

RUN THE SIX-QUESTION PASS
1. Start with relevance. Audit all six questions separately. Check that the item is actually a women's t-shirt, then review fit, fabric, construction, measurements, sizing, care, color, and any documented benefit before adapting copy.
2. Quote the facts used. Mark each connection Supported, Partially supported, or Reject.
3. Supported only: choose one evidence-backed angle around everyday layering, casual outings, and spring and summer timing, then write exactly four assets: Product page or FAQ; Facebook or Instagram; 30-to-45-second short video with shot plan; X post targeted to 240 characters and never over 280.
4. Attach exact facts used to every asset. Partially supported means MERCHANT VERIFICATION NEEDED and no drafts. Reject means explain the conflict and create no drafts.
5. Do not publish, schedule, or present any draft as approved.

CLAIM BOUNDARIES
- Use supplied product evidence for every claim; buyer examples are not measured demand.
- Treat category counts and labels as dated evidence, not demand or performance metrics.
- Use benefits, settings, seasons, and audiences only when individual product facts support them.
- Keep garment identity separate from similar titles, neighboring categories, and example buyer language.

RETURN
Return one table with one row for each of the six questions above and these columns: Question; Decision (Supported, Partially supported, or Reject); Facts used (quoted from the product record); Missing facts; Drafts (all four for a Supported row, none otherwise); Asset status (Publish, Verify, Revise, or Reject for each draft). Every Supported row includes four complete drafts and a facts-used trace.