suede-campaign-in-a-box
Suede AI Campaign in a Box Skills. Use when asked to "plan a release campaign" for a song, era, show or catalog: select a lane and build announce, tease, launch and follow-up assets with rights gaps and fan actions. NOT FOR: isolated copy (suede-copy), sync packages (suede-sync-packaging), rights resolution (suede-rights-passport), or software releases (suede-launch-packaging).
- License 许可
- Apache-2.0
- 最近更新
- 2026-11-13
- Stars
- ★ 1.6k
- Forks
- 483
01 前言 Front Matter
- name
- suede-campaign-in-a-box
- metadata
- { "version": "1.0.0", "display-name": "Suede AI Campaign in a Box Skills" }
- description
- Suede AI Campaign in a Box Skills. Use when asked to "plan a release campaign" for a song, era, show or catalog: select a lane and build announce, tease, launch and follow-up assets with rights gaps and fan actions. NOT FOR: isolated copy (suede-copy), sync packages (suede-sync-packaging), rights resolution (suede-rights-passport), or software releases (suede-launch-packaging).
02 资源文件 4
- 参考文档agents/openai.yaml619 B
- 参考文档CARD.md4.6 KB
- 参考文档references/lanes.md11.7 KB
- 参考文档references/worked-contract.md1.9 KB
03 SKILL.md 正文
Suede AI Campaign in a Box Skills
Read the worked delivery contract for a concrete input, expected artifact, failure response and validation example.
Compatibility
Works from artist/release context, rights records and available media; file tools for packages and calendar drafts. Rendering or publication requires the relevant existing tools and separately authorized accounts.
Gate policy
Every check and verdict here is advice to the user, not a control on the agent. Run the checks, report results honestly, and complete what the user asked. A failed gate changes the report, never the work. The one exception: a finding that risks data loss, credential exposure, a legal or rights violation, a payment mistake, or irreversible public damage. Pause, name the risk, give the options, and let the user choose.
The all-in-one artist campaign skill: it turns a song, release, era, catalog moment, show, or drop into a complete campaign an artist can actually execute. Every campaign capability lives here as a labeled lane. Pick the lane you need, or run the full package that stitches them together.
Core principle: you organize and prepare. You do not clear rights, confirm ownership, approve payouts, write to any registry, secure placements, manufacture cosigns, or guarantee that anything will go viral or sell. Say what to test; never promise the result.
Pick the lane (router)
Read the request and route to the lane that fits. You can chain lanes: most real campaigns use several. If the request is "do the whole rollout," run Lane 0 (Full Campaign Package) and pull the other lanes in as sections.
| If the user wants... | Go to lane |
|---|---|
| The whole rollout packaged: announce, teaser, release week, post-drop, fan proof, catalog afterlife | 0. Full Campaign Package |
| The artist to feel recognizable before any bio/campaign/site is written, identity, world, voice, anti-references, fan promise | 1, Identity Forge |
| A release turned into a recognizable world: visual language, symbols, color, wardrobe, content behavior, rollout tone | 2. Era Builder |
| One track expanded into a creative universe: scenes, characters, imagery, captions, drop ideas, mechanics | 3. Song To Universe |
| The 5-15 second moments that people clip, duet, remix, chant, or share | 4. Hook Hunter |
| A memorable launch moment: stunt, fan mission, puzzle, timed drop, geo clue, unlock, collector action, street-team mechanic | 5. Release Stunt Lab |
| Repeatable fan behavior: phrases, gestures, comments, unlock actions, live moments, collector rituals, trackable tasks | 6. Fan Rituals |
| Visual direction: visualizer, lyric video, canvas loop, stage loop, cover motion, AI video prompts, teaser edits, scene boards | 7. Visualizer Director |
| Merch and physical/collector objects beyond generic logo shirts, tied to lyrics, lore, and fan behavior | 8. Merch Object Lab |
| A live set shaped like a show: arc, intros, transitions, crowd moments, visual cues, encore, talk breaks, merch tie-ins | 9. Setlist Theater |
| Old songs, demos, takes, covers, stems, live clips, unreleased folders, anniversaries, or forgotten assets revived | 10. Catalog Resurrection |
| Collaborator, remixer, producer, visual artist, venue, brand, or creator matchmaking with outreach angles | 11. Collab Matchmaker |
Default starting order when identity is fuzzy: Lane 1 → Lane 2 → (Lane 3 if it is one song) → hooks/stunts/rituals/visuals/merch/live as needed → Lane 0 to package. If the artist already has a clear world, jump straight to the lane the request names.
Multi-agent vs single-agent (ask up front)
This skill can run as a coordinated multi-agent team: one agent per lane, plus a packager that reconciles them into one campaign. Before spawning any fleet, ASK: "Run this as a multi-agent team (more thorough, may use more tokens) or as a single agent (faster, one pass)?" Never silently spawn a fleet. If the user does not choose, default to single-agent and offer to escalate. In multi-agent mode, keep one shared identity/era spine so the lanes do not contradict each other, and have the packager resolve conflicts before output.
Model and cap, state both on the same ask. A spawned agent inherits the session model unless the dispatch names one, so name the model explicitly on every lane dispatch (do not let it default). Cap concurrent lane agents at 4. Twelve lanes plus a packager is 13 agents; run them in waves of 4 or fewer, or ask first. Going past 4 requires naming the model and the rough cost in the same question and getting an answer before launching: an inherited session model is not an answer.
Lane Playbooks
The twelve lane playbooks are in references/lanes.md. Pick the lane with the
router above, then read only that lane's section. The Public Copy Gate and
evidence boundaries below apply to every lane and stay here.
Public Copy Gate (applies to every lane)
Before outputting captions, emails, DMs, press angles, site copy, bios, one
sheets, CTAs, or pitch language, run the Suede AI anti-slop line edit (suede-deslop). Name the
actor, preserve the concrete artist/release artifact, cut throat-clearing,
negative listing, fake intensity, lazy extremes, passive actor-hiding,
pull-quote slogans, generic AI phrasing, unsupported claims, and em dashes.
Evidence boundaries (non-negotiable, every lane)
These lanes organize and prepare campaign material. They do NOT:
- clear rights, confirm ownership, or resolve sample/contributor/clearance questions: flag stale or uncertain rights, samples, contributors, and likeness instead of asserting they are cleared;
- confirm ownership or write anything to a registry;
- approve, route, or guarantee payouts, payments, or fulfillment;
- secure placements, sync, endorsements, partnerships, or cosigns: never imply partnership, endorsement, or access that is not confirmed;
- invent streams, press, traction, biography, or cultural status;
- promise virality, sales, or any outcome: say why something might work and what to test;
- use fake hype, fake scarcity, manipulative claims, or unsafe fan behavior;
- copy another artist's protected identity or assets; no competitor product names.
Branded visual output (Lane 7 visualizers, cover motion, teaser edits; Lane 8
merch objects) uses only the approved Suede S mark file named in
suede-launch-packaging: never redraw, trace, recolor, or generate a
replacement. If that file is unavailable, block the branded visual and request
it rather than substituting artwork.
Never resolve a rights question in-lane. When any lane touches ownership,
samples, contributors, splits, likeness, or clearance, mark the item UNKNOWN or
UNCONFIRMED and route it to suede-rights-passport (audit mode): the campaign plans around
the gap, never over it. When facts are unknown, mark them unknown. Keep safety,
venue, privacy, payment, and platform-rule constraints visible in the output.
Red flags: stop
If any of these appear in your reasoning, stop and re-read the evidence boundaries:
- "The artist says the sample is cleared." A claim is not clearance. Mark it
UNCONFIRMED and route to
suede-rights-passport. - "Imply the collab or placement is confirmed so the outreach lands harder."
- "Say 'only 500 made'; scarcity sells." Not unless the cap is real.
- "Write 'as seen in'; press will probably come." Invented traction.
- "Promise this hook goes viral; it scores high." Say why it might work and what to test. Never promise the result.
How to close
End with the relevant lane Output block(s). When the work produces a final explanation the artist will act on (a full campaign package, a pitch), lead with a Simple explanation (plain, for a 10-year-old): one plain paragraph covering what this campaign is, what the fan is asked to do, and what happens next; no jargon, no hype, no industry words. Then give the normal breakdown: the lane Output blocks, the build-next CTA, and any flagged rights/safety gaps.
Routing
- Rights, sample, split, AI-authorship, or clearance gaps surfaced in any lane → suede-rights-passport (audit mode, then package).
- Release page, artist site, or lyric/credits page needs to rank in search or get cited in AI answers → suede-seo-audit for the page, suede-ai-seo for AI-answer visibility.
- An app tie-in (lesson app, companion app) needs a store listing → suede-aso.
- Track pitched for film/TV/ads/games → suede-sync-packaging.
- Release folder and metadata readiness → suede-release-linter.
- Standalone conversion copy outside this campaign → suede-copy.
04 同仓库更多 Skills
migrate-to-e2e
Moves an existing Playwright, Cypress, Selenium, Appium, Detox, Maestro, or TesterArmy suite to e2e one slice at a time, with the old suite running beside the new one until each ported test is proven. Keeps a MIGRATION.md progress file so any later session resumes where the last stopped. Optionally plants a defect to show both suites catch it. Use for /migrate-to-e2e, "migrate our tests to e2e", "port these Maestro flows", "move our TesterArmy tests", or when a project has an e2e config and an older suite still running.
e2e
Writes, runs, and debugs end-to-end tests for web and mobile apps with e2e, the agentic test runner, which mixes agent.act goals with exact locator assertions on a Playwright browser or an iOS or Android device. Also drives a project's running app live over the e2e MCP server to look at a screen, check a UI change, take a screenshot or record a video, and try a locator; and runs bug bashes, parallel e2e explore runs with each bug proven by a repro test. Use when a project depends on e2e or has an e2e.config.ts, when asked for end-to-end, browser, mobile, or agentic UI tests, to check a change in the running app, to QA or bug bash an app or branch, or when an e2e run fails.
create-verification-skill
Generate a project-local verify-<app> skill on top of e2e, so any coding agent can launch the app, check it is worth driving, exercise a feature the way a user does, keep evidence, and bug bash it. Writes the skill, a feature map with one explore charter per feature, and proves it once. Use for /create-verification-skill, "make a verification skill for this repo", or when a project has no scripted way to prove UI behavior.
deja-search
deja-vu memory — search the user's past AI coding sessions with the deja CLI. Holds what they already settled and never wrote into the repo (naming rules, banned or already-tried tools, required checks before merge, preferred values, who owns what). Use before your first edit in a task, before you call a change done or ready to merge, before you recommend a tool, library, design or reviewer, before re-debugging an error that may already be solved, when they say 'didn't we fix this before', 'what did we decide about X' or 'I already have that', and before telling them something on this machine does not exist.
deja-reflect
End-of-session reflection and memory curation with deja-vu. Review what was decided, what commands broke and what repaired them, and record durable project rules or decisions with deja remember so the next session starts with them.
revid
Make, edit, export and post AI videos with Revid (revid.ai) through the Revid MCP server. Use whenever the user wants a short-form or long-form video made (from an idea, a script, a URL, an article, a PDF, a tweet, a product page, their own clips, or a song), wants to find what is going viral on TikTok, YouTube, Instagram or X before making something, wants to edit or re-export an existing Revid project, generate a standalone AI image or clip, make a series episode, or publish and schedule a finished video to connected social accounts. Covers setup (hosted MCP with OAuth, or an API key), the 36 tools, credits, polling and the rules for publishing. Routes to the more specific revid-* skills when one fits.