Skip to content

gooseworks-ai/goose-skills / skills/marketing/work-with-creators

work-with-creators

Verified

Real creators and influencers, researched, priced and briefed. Use when the marketer says "find influencers", "which creators are worth paying, at what price?", "find real UGC creators to film for our ads", "who should we seed product to", "is this creator's engagement real", "what should I pay this creator" or "write a creator brief". Free research first; one credit yes only to vet deeper. Drafts only: contacts nobody, agrees nothing, moves no money. Not for AI-made UGC-style video — use goose-video; if unsure, ask: a real person filming, or AI-made?

references/creator-brief.md

View on GitHub
# The one-page creator brief

Load at step 7. One brief per creator — a generic brief sent to everyone is the commonest failure. The brief tells the creator what must be true and what must not be said, then gets out of the way: over-scripted content dies.

## Sections (in this order, one page)

1. **Header** — creator, brand, product, campaign (if any), go-live window.
2. **Why you** — one specific line about a past post of theirs (shows the brief was written for them).
3. **Goal and audience** — one line each, in plain words.
4. **Key message** — one sentence, built only from signed-off claims or neutral facts (the kit's value props and the brain's unsigned "approved" claims are context, not claims).
5. **Product facts** — at most three, **neutral** facts from the product records: format, size, key ingredients, how to use, price (country of origin is a regulated claim, not a neutral fact). SaaS and apps: features, integrations, plan price from the product records; none on file → ask the human for 2–3. Product-page copy often carries efficacy language ("fights breakouts", "kills bacteria") — that is not a signed-off claim and does not go in a brief.
6. **Hook angles** — 2–3 *angles* written for this creator's format and style (a stitch, a routine, a "things I wish I knew"). Not scripts, not lines to read. Same claims rule as the key message.
7. **Must-says** — at most three, each a **signed-off claim** in its exact words, cleared through the `claims-gate` skill (`../../claims-gate/SKILL.md`) (it reads the sign-off; with none, it puts the few lines this brief needs to the human for a yes). Lines the human doesn't approve are cut and listed with what proof would restore them; the brief then reads "not ready until you approve: <lines>" only if its message depends on them. **Can cite (optional):** press quotes only when they are among the signed-off claims — quoted as the publication's words, never as the creator's experience. Ratings and review counts are claims: signed-off and current, or cut. Policies (shipping, returns) only from the brand's `must` facts.
8. **Never-says** — claims that are not signed off (in regulated categories the creator's **own results** are claims too — "it cleared my breakouts" is an acne drug claim: signed off or cut; acne words: clears / treats / prevents acne or breakouts), the brand's banned words, disease/medical words for regulated categories, competitor names, anything the brand's operating rules forbid.
9. **Mandatories** — disclosure (below), handles to tag, `[LINK]` / `[CODE]` placeholders the human fills when deals are agreed.
10. **Deliverables and specs** — platform, format, length, aspect ratio, count, posting window, minimum time live. **Newsletters:** placement (primary, secondary or dedicated send), word count, who writes the copy (usually the writer, from talking points), send date, tracked `[LINK]`. **Podcasts:** host-read talking points (not a script), pre- or mid-roll, length, promo URL or code placeholder.
11. **Timeline** — product arrives (if gifted), draft due, feedback within ~48h, live date. Realistic: rush costs extra.
12. **Approval** — one round, covering claims and disclosure only.
13. **Usage rights (proposed, pending contract)** — organic repost term; paid usage / whitelisting / partnership-ad access and term; exclusivity window if wanted. The human negotiates; the brief only states the ask.
14. **What not to do** — 2–4 lines (e.g. "don't read the product facts to camera", "no before/after claims").
15. **Creative freedom** — "Your voice, your edit. We only check claims and disclosure."
16. **Two examples** — (a) one of **this creator's own** posts that already fits the brief; (b) one category winner for the *format* — prefer a creator's post that does not feature a competitor's product; never a brand ad to copy.
17. **Contact** — `[placeholder — the human fills this in]`.

## Seeding variant (gift-only, no obligation)

A gifted creator owes nothing; a paid-deal brief in a $20 box reads as a demand and gets ignored. Keep it to a short note:
- **Why you** — one specific line about their content.
- **What's in the box and how to use it** — neutral facts.
- **If you post** — 1–2 angle ideas in their format; no required deliverable, no draft, no approval round, no deadline.
- **Disclosure if you post** — "#gifted" (or "thanks to <brand> for sending") in the first line and said in the video; the platform's label where it applies.
- **Tag** — the brand's handles.
- **Never-says** — the claims the creator must not make for the brand (disease words, anything not signed off).
- **Rights** — only permission to repost on the brand's own accounts with credit; paid usage is negotiated later if the post performs.

## UGC-for-ads variant (the brand owns, edits and runs the asset)

Use this shape instead of sections 6, 10, 12 and 15 when the role is UGC-for-ads (keep 11, the timeline, and add when the product or app access arrives — before filming):
- **Concepts and hooks** — 1–2 concepts, each with **2–3 hook variants as separate deliverables** (the brand tests hooks).
- **Shot list** — e.g. face-to-camera open, product or screen-recording close-ups, the moment of use, a reaction; lighting and sound basics.
- **CTA** — the ask at the end, in plain words.
- **Deliverables** — finished cuts per hook in 9:16 (and 4:5/1:1 if asked), plus **raw footage** (priced in the deal), no music baked in if the brand adds its own.
- **Revisions** — one round on the creative itself, not only claims and disclosure.
- **Freedom statement** — "Be yourself on camera; we edit from your raw footage."
Everything else (must-says, never-says, honest-use line, rights ask, timeline, examples) stays. **Hook variants and the CTA are lines said on camera in the brand's own ads** — the same claims rule applies to every one of them: signed-off claims or neutral facts only.

## Disclosure (state the requirement; the human owns legal sign-off)

**UGC-for-ads variant**: when the asset runs from the brand's own ad account, the ad is the brand's and is labelled as an ad by the platform — the creator does not add #ad to a post they never publish — but if the brand's ad presents the creator as an ordinary customer, US endorsement rules may require the paid connection to be disclosed in the ad itself: flag it for the human or legal. When it runs from the creator's handle (Spark / partnership ads, whitelisting), the platform's branded-content or partnership label applies. Either way, the creator's words must reflect honest use: no claimed usage duration ("I've used it for months") or results they have not had; for a new product or app, they use it before filming.

The platform and market rules (paid-partnership labels, where "#ad" / "#gifted" goes, spoken disclosure, LinkedIn, newsletters, podcasts, US and UK rules) come from the shared `../../ad-policy-check/disclosure.md` (in the `ad-policy-check` skill) — load it and copy the lines for this creator's platform and market into **Mandatories**. Brief-specific additions: gifted content needs disclosure too; LinkedIn rights ask = permission to run the post as a Thought Leader ad; markets the shared file doesn't cover → "check local rules", never a guess.

## Example (FICTIONAL brand and creator — the bar, not data)

> **Tailwag Joint Chews × @fictional_corgimom** (TikTok, 38k) · live Nov 14–20
> **Why you:** your "senior corgi on the stairs" video — your audience is people with ageing dogs.
> **Goal:** awareness among owners of dogs 7+. **Key message:** an easy daily chew that supports healthy joints in older dogs.
> **Facts:** chicken-flavoured soft chew · one a day by weight · 60 chews a tub.
> **Hook angles:** Biscuit's morning routine, then the chew as one beat in it · "things I wish I'd started when she turned 7" · stitch a vet-tips video with your own experience.
> **Must-say:** "supports healthy joints" · "vet-formulated" (both signed off).
> **Never-say:** "cures", "pain relief", "arthritis treatment", any rival brand.
> **Mandatories:** paid-partnership toggle + #ad in line one + said in the first 3 seconds · tag @tailwag · [LINK].
> **Deliverable:** 1 vertical video, 30–60s, live ≥30 days. **Timeline:** chews arrive Nov 3 · draft Nov 10 · our notes within 48h · live Nov 14–20.
> **Approval:** one round, claims and disclosure only. **Usage (proposed, pending contract):** 30-day organic repost; 30-day Spark Ads access.
> **Don't:** read the facts to camera; show a before/after mobility claim.
> **Freedom:** your voice, your edit. **Examples:** your "stairs" video [url]; a top category post for format [url].
> **Contact:** [placeholder]

## Why briefs fail

Scripts; more than three must-says; creators presenting long use or results they haven't had; feature dumps; one brief for everyone; claims that are not signed off; asking for the brand's voice instead of the creator's; vague deliverables or no posting window; no disclosure line; unrealistic turnaround; several feedback rounds; no rights ask (then the content can't be reused in ads).
add-product-e2e-evalpaperclipai/paperclipAdd or extend a Paperclip full-stack runner E2E workflow, fixture, matcher, or report evidence path for local or Daytona execution.Stars: 115kUpdated Claude CodeCodexVerifiedInstall
add-runner-evalpaperclipai/paperclipAdd or extend a Paperclip Runner protocol evaluation definition, roster, assertion, or report fixture with provenance and narrow validation.Stars: 115kUpdated Claude CodeCodexVerifiedInstall
diagnose-why-work-stoppedpaperclipai/paperclipCompatibility entry point for failed, stalled, looping, or over-recovered Paperclip work. Use fix-failed-paperclip-run for the full diagnosis, regression, repair, real verification, and green-PR workflow; honor an explicitly requested diagnosis-only scope.Stars: 115kUpdated Claude CodeCodexVerifiedInstall
fix-failed-paperclip-runpaperclipai/paperclipDiagnose and fix failed, stuck, slow, or incorrectly blocked Paperclip tasks and runs. Use with a task/run URL or failure report to trace the root cause, prove a regression red then green, check second-order effects, verify the real workflow, and deliver a PR with passing checks.Stars: 115kUpdated Claude CodeCodexVerifiedInstall