菜单
Write conversion copy that earns the click: landing sections, email, microcopy, buttons, headlines, CTAs, variants, and anti-slop edits.
用 Codex 或 Claude 帮你安装 复制这段 Prompt，粘贴到 Codex、Claude 或其他助手里，让它检查 Skill 页面并帮你完成安装。
基于 SOC 职业分类
The pack's contract/customer-service negotiator, proven outside a repo: scans an Amazon account for restocking fees, short refunds, and forgotten or overpriced digital subscriptions (Prime Video Channels like Britbox/Starz/AMC+/Paramount+, Audible membership, Kindle Unlimited, Prime itself) — money Amazon is quietly holding or billing that the owner never noticed — then drives Amazon's live chat to get fees waived, refunds issued, or unused subscriptions canceled and the last charge refunded. Recovered $448.31 in one sitting — including a full refund on an item Amazon had already denied once, with no return required. Use this whenever the user mentions Amazon returns, restocking fees, an Amazon refund that looks short, a forgotten subscription (Britbox, Starz, Audible, Kindle Unlimited, etc.), disputing an Amazon charge, checking whether a return was fully refunded, auditing recurring Amazon charges, or asks something like 'did I get charged for that return', 'am I still paying for Britbox', or 'is Amazon sti
The pack's contract negotiator, generalized beyond Amazon: a recurring-charge auditor that finds forgotten, unused, or overpriced subscriptions across any service (Netflix, Spotify, Hulu, Disney+, gym memberships, SaaS tools, news sites, app subscriptions) and either cancels them directly or negotiates a refund/waiver through that service's own support channel. Complements amazon-returns-recovery, which stays scoped to Amazon returns/restocking fees and Amazon-billed subscriptions (Prime Video Channels, Audible, Kindle Unlimited) — this skill covers everything billed outside Amazon: direct-bill streaming and software, App Store and Google Play subscriptions, and PayPal-billed recurring payments. Use this whenever the user wants to audit recurring charges generally, asks what subscriptions they're paying for, mentions a specific non-Amazon subscription (Netflix, Spotify, a gym, a SaaS tool, etc.), wants to find and cancel unused subscriptions, or asks something like 'what am I still paying for', 'find my subsc
Umbrella workflow for 25 public skills: copy, design, code review, SEO, launch packaging, MCP QA, iOS conversion, and creator workflows.
Claude-directed parallel OpenAI Codex CLI worker fleet for bulk generation. Use when a job is high-volume, well-specified, and splits into independent worker-sized tasks (content batches, test generation, bulk refactors) and Codex CLI is installed and logged in. Claude decomposes, briefs, spawns codex exec runs in parallel, and review-gates every output. NOT FOR: multi-lane Claude agents coordinating one complex change (use suede-agent-teams); low-volume, judgment-dense copy Claude should write itself (use suede-copy or johnny-suede-write).
Design and write polished product surfaces people understand fast: landing pages, dashboards, campaigns, restyles, UI copy, and visual QA.
Make Suede interfaces feel intentional: tokens, color, components, type, motion, dark mode, and visual QA for shipped screens.
| name | suede-copy |
| description | Write conversion copy that earns the click: landing sections, email, microcopy, buttons, headlines, CTAs, variants, and anti-slop edits. |
Write conversion copy, page copy, GitHub docs, email, and social posts that are specific, proof-backed, and free of AI boilerplate. Default voice: Suede. Supply a company brief to override everything.
Core principle: every claim is verifiable or it gets cut, and nothing ships below its score threshold.
Supply a brief and all copy, voice, and claim logic applies to your company. Use natural language or this form:
Company:
Product or offer:
Audience:
Voice:
Terms to use:
Terms to avoid:
Proof:
Allowed claims:
Forbidden claims:
Primary CTA:
Read any available context files before asking questions: PRODUCT.md, README.md, AGENTS.md, AI_HANDOFF.md, DESIGN.md, product marketing or brand notes, task-specific docs.
If context is missing after reading, ask only for what blocks accurate copy:
Name the outcome, not the feature.
Write buttons as actions with a result.
Replace vague claims with artifacts.
No invented proof. Do not write stats, testimonials, partner names, pricing, or legal clearance that has not been confirmed. If proof is unavailable, write around the gap or flag it for the human to supply.
No em dashes. No exclamation points. No rhetorical questions that answer themselves.
Match framework to surface and reader temperature:
State the chosen framework and reader temperature before drafting. If multiple frameworks could apply, pick one and note why.
Generate 3 headline candidates minimum for any hero or email subject. Pick the formula that matches the reader's state and the page's job.
| # | Formula | Structure | Example |
|---|---|---|---|
| 1 | Curiosity gap | [Intriguing partial claim] | "Most release folders fail the first licensing check. Here's why." |
| 2 | Number-led specificity | [#] [specific thing] [timeframe/condition] | "12 rights fields missing from your release. Suede finds them in 60 seconds." |
| 3 | How-to outcome | How to [achieve outcome] [without/with condition] | "How to package a release folder that licensing teams can actually use" |
| 4 | Because | [Result] because [mechanism] | "Agents can read your music rights because Suede structures the provenance first." |
| 5 | Specificity anchor | [Exact number or name] + [claim] | "47 fields. One linter. No guessing." |
| 6 | Before-After | [Before state] → [After state] | "Scattered files and a split sheet in a Google Doc → a machine-readable rights package" |
| 7 | Question (real, not rhetorical) | [Question reader actually asks] | "What does a licensing team check before they sign?" |
| 8 | Objection flip | [Common objection] + [reframe] | "Rights metadata sounds like legal work. It's a 10-minute audit." |
| 9 | If-then conditional | If [specific situation], then [specific outcome] | "If your release goes to a sync library, this is the metadata they'll reject first." |
| 10 | Direct claim with proof hook | [Bold claim] + [verifiable detail] | "Suede reads your folder. You get ISRC, split, and flags before you pitch." |
| 11 | Problem named exactly | [Specific failure mode the reader fears] | "Your ISRC is assigned. Your split sheet is a PDF. Neither is machine-readable." |
| 12 | Authority + specificity | [Who trusts this] + [for what exact task] | "The metadata structure sync licensing teams check on day one." |
Test: swap your product name for a competitor's. If the headline still works, it is not specific enough.
State the persona before writing. If multiple personas share a page, write the hero for the decision-maker and include practitioner proof in the secondary section.
Decision-maker (exec, founder, buyer)
Practitioner (developer, designer, operator)
Skeptic (comparison shopper, previously burned)
Creator / end-user (non-technical)
For a page, README, or docs surface, build this spine:
For a small section, use only the pieces that fit.
For high-stakes copy (hero headline, primary CTA, email subject, ad copy), always generate variants.
Headlines: 3 variants, different angles:
CTAs: 2 variants minimum. See CTA Formulas below.
Email subjects: 3 variants:
Label each variant with its angle. Let the user pick rather than guessing.
CTAs fail when they describe the button, not the outcome. Every CTA answers: "What happens the moment I click this?"
Formula A: Verb + immediate result [Action verb] + [what they get or see right now]
Formula B: Verb + object + benefit [Verb] + [specific object] + [value unlocked]
Formula C: Low-commitment framing (skeptic / discovery stage) [Passive discovery verb] + [what they'll see, not what they'll do]
Formula D: Stakes-aware framing (decision-maker) [Verb] + [outcome in their language]
Anti-patterns to cut:
The 3-word test: describe what happens after clicking in 3 words. If you cannot, the CTA is too vague.
Subject lines win opens on three mechanics: curiosity, self-interest, or specificity. Pick one per subject line.
Curiosity:
Self-interest:
Specificity anchor:
Avoid:
Preview text is a second subject line. Write it to add information, not echo the subject.
Keep preview text under 90 characters. If the client truncates at 40, the first 40 characters must stand alone.
Hook (1-2 sentences): Name the problem or opportunity at the exact moment the reader is experiencing it.
Proof or evidence (2-4 sentences): Specific, not general. One example beats three claims.
Bridge (1 sentence): Connect the proof to the offer.
CTA (1 sentence + link): One action, one link. No secondary options in the primary CTA block.
P.S. (optional): Use for a single secondary offer or a time constraint. Not both.
Unsubscribe-reduction:
Hook line (first 2 lines before "see more"): The hook is the post. If the first two lines do not earn the click to expand, the rest does not matter.
Post structure:
[Hook: one specific claim, observation, or question]
[2-4 line break]
[Insight or story: 3-6 short paragraphs, one idea each]
[Takeaway: what the reader does with this]
[CTA: one, low-friction. "What's your experience?" or a link, not both]
Formatting rules:
Standalone tweet:
[Specific observation or fact] + [one implication or action]
Max 240 characters. If it reads like a self-contained thought from someone who knows something, it is working.
Thread opener:
[Bold specific claim]
[Thread: number + what the reader gets]
"Here's how it works, step by step:"
The thread opener must be the strongest tweet in the thread. Do not save the best point for tweet 5.
Reply to trend or news:
[Acknowledge the news in 1 sentence]
[Specific take from your vantage point]
[Optional: link to your related resource]
Caption structure by content type:
Product / feature reveal:
[Name what it does in one sentence (the hook)]
[Why that matters for this specific audience]
[One specific proof point or use case]
[CTA in bio or link sticker]
Behind-the-scenes / process:
[Name the specific moment or decision shown]
[What you learned or chose and why]
[Invitation: "What would you have done?"]
Testimonial / social proof:
[Lead with the result, not the quote]
[Quote or paraphrase the proof]
[Bridge to your offer]
[CTA]
Caption rules:
Use this register: confident, not breathless; technical enough for builders; clear enough for creators; polished, not corporate; specific, not cute; operator-grade, not brochure-grade.
Good Suede copy names what the reader controls: register a work, verify rights, route royalties, publish a claim, package a release folder, prepare licensing evidence, make a work readable to agents, compare provenance, ship a public skill page.
(For non-Suede work, supply the equivalent domain vocabulary in the company brief.)
For GitHub repositories, skill docs, and Pages sites, treat SEO as the umbrella for search, AEO, and AI EO. Include:
Suede durable keywords: Suede Creator Skills, Suede Rights Passport, Music Release Metadata Linter, Suedify, Suede Copy, AI EO, AEO, answer engine optimization, Codex skills, Claude Code skills, SKILL.md, music rights, creator rights, release readiness, provenance, royalty splits, licensing readiness, programmable IP, agent commerce, GitHub Pages.
Use keywords because they help the right reader find the page. Do not cram a keyword where a human would notice.
For a deep, standalone SEO audit (technical access, keyword research, schema markup, E-E-A-T signals, topic cluster architecture, AI EO optimization, and scored visibility grades), use $suede-seo-audit instead.
When the copy workflow includes an SEO pass (metadata, structure, or copy quality only):
Run this as a line-edit gate before delivery, not a vibe check.
Make these swaps. Non-negotiable.
| Cut | Replace with |
|---|---|
| utilize | use |
| leverage (as verb) | use, apply, run |
| seamless | remove or prove it: "no export step", "one command" |
| powerful | prove it: "processes 10k records in 4 seconds" |
| innovative | cut; name the innovation instead |
| revolutionary | cut entirely |
| game-changing | cut entirely |
| solution | name what it actually does |
| ecosystem | platform, system, toolchain (pick the accurate one) |
| empower | cut; say what the person now controls |
| unlock | cut; say what was blocked and is now accessible |
| streamline | speed up, cut the step, reduce from X to Y |
| intuitive | cut; prove it with a UX detail |
| robust | cut; name the specific capability |
| best-in-class | cut; or supply the benchmark |
| next-generation | cut; name what changed |
| cutting-edge | cut entirely |
| world-class | cut; or name the credential |
| end-to-end | name the actual start and end |
| holistic | cut; describe the scope instead |
| scalable | prove it: "handles X at Y load" |
| simple / simply / just / easy | cut or prove it with a step count |
| we believe / we think | cut; make the assertion directly |
| in order to | to |
| due to the fact that | because |
| at this point in time | now |
| going forward | cut; state the new behavior directly |
| synergy / synergistic | cut entirely |
| value-add | name the value |
| best practices | name the practice |
Aim for Flesch-Kincaid Grade Level 8-10 for B2B general audiences, Grade 6-8 for consumer audiences and onboarding, Grade 10-14 for technical/developer audiences where precision requires complexity.
Average sentence length under 18 words for consumer; under 22 words for B2B. Flag paragraphs over 4 sentences.
Rewrite:
Name who does the action. Prefer the creator, operator, buyer, agent, page, repo, workflow, file, command, route, or proof artifact.
The page converts traffic.The page routes visitors to the audit, the proof link, or the build request.The market rewards provenance.Licensing teams can inspect the provenance trail before they ask for the split sheet.always, never, everything, and nothing unless the claim is literally true.If a line sounds manufactured for a quote card, rewrite it with a real artifact, action, or proof point.
The future of creator ownership is here.Suede turns a release folder into rights, provenance, split, and licensing evidence an agent can read.We're changing how music rights work.Paste a folder path. Suede returns your ISRC status, missing fields, and a split-ready JSON file.Built for the next generation of creators.A sync licensing team can open your release folder and read every rights claim without calling you.Score the copy before handoff:
Directness: /10
Rhythm: /10
Trust: /10
Specificity: /10
Authenticity: /10
Density: /10
Search/AI readability: /10
Total: /70
Revise below 58/70. For public launch, homepage, GitHub, App Store, investor-adjacent, or public explainer copy, aim for 62/70 or higher.
Title:
Meta description:
Hero:
Subhead:
Primary CTA:
Sections:
FAQ:
Final CTA:
Safety note:
Skill:
One-line description:
Reader:
Primary action:
Repo/Docs copy:
Install CTA:
SEO title:
Meta description:
Keywords:
Safety boundary:
Findings:
Rewrites:
Claims to verify:
Score:
Ready: yes | with caveats | no
If any of these thoughts appear, stop and run the gate you were about to skip:
Do not ship copy when:
End with the exact copy, not a long explanation of the copy.
Accept feedback at any point, not only after final handoff. When the user says what worked, preserve that pattern in the current pass and mirror it later. When the user says what missed, adjust immediately instead of defending the previous direction.
If the user says cue suede, asks for feedback choices, or seems to be calibrating mid-stream, pause at the next safe checkpoint and offer:
Cue Suede:
1. Change something - tell me what to revise and I will adjust it.
2. Preserve this - tell me what worked so I can mimic it later.
3. Keep as-is - say nothing and I will treat it as accepted.
Do not block completion waiting for a Cue Suede answer. If the interface supports choice chips, use Change something, Preserve this, and Keep as-is.