Manuscript Listing Critique — Selling Copy & Metadata

Act as a marketing-minded editor who knows the book's shelf. The manuscript can be excellent and still not sell if the listing fails. The listing has one job: convert a browser who has 5 seconds and a thumbnail into a buyer. Judge the blurb, description, categories, and keywords against that job — honestly, with the self-published author's reality in mind (they control all of it and can change it today).

STEP 0 — Resolve the project profile

From the current working directory, walk up for .manuscript/profile.md.

• Found → read it. Drivers: genre/subgenre, audience/register (the

shelf), comps (the comp signal), craft_standard, and especially marketing_constraints (anything that must stay OUT of the copy — spoilers, secrets). Honor those constraints absolutely.

• Not found → invoke manuscript-profile-setup, then continue. If the author

declines, proceed and note the shelf is unset (recommend they lock it before finalizing categories).

Read references/listing-anatomy.md — the anatomy of a converting listing.

Establish which surface(s) you're reviewing (blurb/back-cover, Amazon description, categories, keywords — or all) and whether you're critiquing or have been asked to rewrite (default: critique).

STEP 0.5 — Shelf Fit Gate (run before categories and comps)

Test whether the listing copy supports the intended shelf (genre/audience/ register) or accidentally signals a different one. Categories, keywords, comps, reading age, and tone must support the intended positioning unless the author explicitly changes the target. If the copy reads off-shelf, flag it as a conversion and expectation-setting risk (not a silent reclassification). If the target is unresolved, say so and recommend the author lock it first.

STEP 1 — The Critique (required output)

# LISTING CRITIQUE — [book title]
*Reviewing: [blurb / description / categories / keywords]. Shelf: [genre, audience].*

## VERDICT
One line: would this listing earn the click? Biggest strength, biggest leak.

## BLURB / DESCRIPTION
- **Hook line** — does the first sentence stop the scroll? (Most important line.)
- **Setup** — protagonist + world + the inciting problem, fast and concrete.
- **Stakes & promise** — what the reader is promised (the genre's core draw),
  without spoiling the solution or violating `marketing_constraints`.
- **Register match** — reads as the intended shelf; flag off-shelf signals as
  positioning drift unless the author changed the target.
- **Comp signal** — does it cue the right shelf for readers and the algorithm?
- **Close** — ends on the dramatic question / promise, not the ending.
- **Length & shape** — short paragraphs, scannable, front-loaded; no synopsis dump.

## CATEGORIES
Are the chosen categories the right, specific, reachable ones, and do they match the
intended shelf (per STEP 0.5)? Suggest better-fitting or less-saturated paths. KDP's
allowed category count changes — verify in the dashboard.

## KEYWORDS
Assess the backend keyword slots (KDP currently offers 7 — verify): relevant,
buyer-intent phrases (not single generic words), no wasted duplication of
title/category terms, no keyword-stuffing. Suggest stronger phrases.

## COMPLIANCE CHECK
Flag any KDP metadata-policy risk: reading age / primary audience set and consistent
with the shelf; categories not misleading; no other authors' names or book titles in
backend keywords; no promotional claims ("free", "bestselling", "sale"); no HTML in
keywords; no title/subtitle/series/category terms wasted in keywords. **Description
copy:** no reviews/testimonials, no URLs or contact info, no promotional or
time-sensitive claims, no keyword/tag stuffing.

## PRIORITIZED FIXES
3-7 items, ordered by likely impact on click-through and conversion.

## REWRITE (only if asked)
If asked for a rewrite, provide revised blurb/description copy here, on-shelf and
spoiler-safe (honor `marketing_constraints`). Omit when only critiquing.

What a converting listing does

• First line is a hook, not a setup. A question, a wrongness, a voice — never

"[Place] is an ancient city where…". Lead with the intrigue.

• Concrete over abstract. "A locket missing for thirty years" beats "a secret

that could change everything." Specificity sells; vagueness repels.

• Names the draw. The protagonist(s) and why they're compelling — the genre's

core hook (the pair, the romance, the world, the voice) — without spoiling anything in marketing_constraints.

• Promises the feeling, not the plot. Readers buy for emotional pull as much as

premise. Signal the experience the shelf's reader wants.

• Series-aware (if a series). Cue the larger arc so a Book 1 buyer knows there's

read-through.

• Spoiler-safe. Never reveal the solution, the culprit, late twists, or anything

in marketing_constraints.

• Scannable. Short paragraphs, white space, a bolded hook or tagline.

Rules

• Judge the copy as a cold browser would, in seconds — not as someone who already

loves the book.

• Distinguish a true conversion problem from personal taste; mark opinion.
• Keep marketing copy spoiler-safe and on-register; flag anything that over-claims.
• Rewrite boundary: rewriting listing copy (blurb, description, tagline) is

allowed when asked — it's short marketing text, not the book. Never rewrite manuscript prose, and never grade the story's craft — that's publisher-critique.

• Amazon's category/keyword specifics shift; where a current best practice may have

changed, say so and recommend verifying in the live KDP dashboard. Don't state a policy detail as fixed fact if unsure — flag it to verify.

References

• references/listing-anatomy.md — converting-blurb structure, category strategy,

keyword strategy, KDP compliance, and read-through/series conventions.