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/discovery.md

View on GitHub
# Discovery

Load at step 2. Discovery is free when done the way a person would: web search, the platforms' public search, hashtag and profile pages in the host browser, rivals' posts and partnership ads. Discover wide for free, screen on the public profile, vet finalists from their public pages; paid reads only widen or deepen the result after the founder's one yes.

## Sources, best first

1. **People already talking about the brand** — the best first list *when the mention is positive and about the product* (for a celebrity-founder brand, posts about the founder are not product fans — a fan shows or reviews the product): they chose the brand unpaid, their audience already trusts the mention, and gift-first works best here. Read each mention's sentiment first — brand-name search also returns "honest review / it broke me out" videos and drama; those are near-misses (or risks to tell the human about), not fans. A brand with no mentions in its free organic feed (new brands) skips the paid fan search.
   - Posts that tag the brand's own accounts (tagged posts on Instagram; brand-name keyword search on TikTok, Reels, YouTube, X, Threads).
   - Creators among the brand's own followers and commenters on its recent posts (the brand's organic feed is free in the brain). For a founder-led or celebrity-founder brand, the founder's own account and the creators who engage with it.
   - Creators the human already knows of — ask, once.
2. **Creators proven in the category**
   - Competitors' recent organic posts that tag or feature a creator (free: the brain's competitor organic feed). Check each is a creator in the category before using it — rivals also tag celebrities, comedy series and kids' channels, which are not candidates for this brand. When a dated public page contradicts the brain (a collab the web can't confirm, a rating that differs), the newer dated page wins; say so in the report and flag the brain item.
   - Competitors' **partnership ads** in the Meta Ad Library, read free in the host browser (a brand paid to amplify them — proven to convert; check conflict).
   - The people behind research-market's saved outliers in the Inspiration library (tag `research-market`) — strong *format* fit; check they are consistent, not one viral hit.
3. **Problem-phrase search** — web search and each platform's own search page in the host browser, on what the persona types or says ("acne safe moisturizer", "dog limping on stairs", "month-end close takes forever"), not just hashtags. Hashtag search over-surfaces big accounts.
4. **Smaller creators in the comments of category leaders** — the micro tier often lives there.
5. **Lookalikes** — collaborators, duets and stitches of good candidates; accounts suggested alongside them; who a strong candidate follows.
6. **"Who do my customers follow"** — only if the brand supplies a customer survey or list. Ask; never guess.

Channel-specific:
- **Newsletters**: web search through the proxy for `site:substack.com` / beehiiv / "sponsor" or "advertise" pages in the niche. Newsletters come back as a candidate list with "size creator-claimed — request the media kit" unless a public badge or advertise page states it.
- **Podcasts**: category podcast search and charts; guest overlap between shows.
- **LinkedIn**: topic post search, then the commenters on top voices (the buyers and the mid-size voices are there).
- **UGC-for-ads creators**: creators whose content appears in brands' partnership ads in the category (ad library), profiles that say "UGC creator" with a portfolio on their link-in-bio, and keyword search for demo-style videos in the niche. (UGC marketplaces are the human's own sourcing route.)
- **YouTube**: search by problem phrase; the "suspected sponsors" check on a candidate's recent videos shows who already pays them.

## The funnel (approximate)

| Stage | Count | Screen on |
|---|---|---|
| Long list | ~50–150 (a fetch-only free pass: ~15–30) | Found by the sources above; one line each: handle, platform, source, why surfaced |
| Quick screen | ~30–80 | On the fields search answers already return, then a profile read for the top ~25 only: posted in the last ~14 days, on-niche, right language/geography, tier inside the plan, not a minor, no obvious safety issue in the bio and top posts, not already in the roster as declined/never-again |
| Finalists | ~15–25 for paid and UGC roles; ~30–60 for a gift-only seeding wave (free pass: 5–12 well-read beats 25 half-read) | Ranked by source strength (fans > proven-in-category > search) and screen notes — this is the list quoted at Gate 2, split into full and light vetting |
| Shortlist | ~5–15 paid/UGC; ~20–50 for a seeding wave | After vetting (full or light) and pricing |

Every creator cut after vetting becomes a near-miss with a one-line reason and the number behind it. Creators cut at the quick screen are near-miss rows with `stage cut = screen`, no extra spend.

## Do not

- Pick on follower count, verified badges or "looks on-brand" from a profile picture.
- Treat one viral outlier as a creator's normal.
- Use a creator the roster marks "never again" or "declined" without saying so.
- Spend vetting credits on anyone who fails the free or profile-level screen.
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