gooseworks-ai/goose-skills / drafts/marketing/learn-my-brand
learn-my-brand
VerifiedLearn my brand once and stop asking me every time. Use when someone says learn my brand, set up my brand, here's our brand book, remember how we talk, what can we claim, can we say this, approve our claims, sign off our voice, our price or product changed, the client approved the claims, pick up where we left off, or check our claims are still current. Reads what GooseWorks already found about the brand, asks only the gaps as drafts to confirm, audits every claim (including ones research marked approved) and gets the ones that matter signed off with proof, scope and review date, turns the voice into rules the owner confirms, checks logo, colours and fonts, and saves it all as it goes so every other job loads it instead of asking. Ends with an unknowns list and a read-back. Not for who our customers are and what they say (use research-audience) or writing positioning from scratch (not covered yet).
| File | Kind | Size |
|---|---|---|
| SKILL.md | skill md | 14.5 KB |
| atoms/brain-write-back/ATOM.md | markdown | 6.8 KB |
| atoms/claims-audit/ATOM.md | markdown | 3.9 KB |
| atoms/voice-calibrate/ATOM.md | markdown | 3.1 KB |
| references/brain-map.md | markdown | 10.2 KB |
| references/brand-record-template.md | markdown | 6.2 KB |
| references/category-rules.md | markdown | 9.5 KB |
| references/claims-craft.md | markdown | 13.4 KB |
| references/intake-craft.md | markdown | 7.8 KB |
| references/voice-craft.md | markdown | 6.4 KB |
| skill.meta.json | config | 4.5 KB |
| tests/test-by-hand.md | markdown | 12.0 KB |
references/intake-craft.md
View on GitHub# Intake craft
Load at steps 2 and 3. How a strategist fills the gaps in week one without turning it into a form.
## Scope: what is settled today and what waits
The founder answers a small set of brand-owned facts now. Everything else is recorded as unknown, with the playbook that will ask for it at the moment it needs it. Page 01's rule: ask about ten facts on day one, never a long form.
**Settled in this playbook (confirm a draft, never ask cold):**
| Day-one fact | What "settled" means |
|---|---|
| What we sell, and the hero offer | The products the brain found are right; which one is the hero; the standing offer, if any; anything sold out or never to advertise (`ad_eligibility: hidden`, with the founder's reason). |
| Who it is for | The founder's own one-line answer, confirmed against the brain's audience read; who it is *not* for. Top objections are not asked here: they come from research-market's customer evidence (ledger: "waits for customer research"). |
| Company type and stage | The brain's `brandType` confirmed; stage in the founder's words (pre-launch, launched, growing, stalled). |
| What we may claim | The claims audit and sign-off (claims-craft.md). |
| What we must never say or do | The never-say list: law, platform and owner entries (settled in step 4, with the one question about what must not be said publicly). |
| Voice examples, logo and colours | Voice sign-off (voice-craft.md); logo, colours and fonts confirmed (step 6). |
| Who approves | Who signs off claims and voice. If it is not the person in the chat, new lines wait for the approver on the sign-off sheet; claims already approved stay approved. |
| One-line positioning, if the founder has one | Their own words: for whom, what it is, instead of what. If they have none, record it as unknown and name write-my-positioning (backlog); do not write positioning here. |
| What buyers use instead of us | One question: who they lose to, or (pre-revenue) what buyers would use instead, including doing nothing. Names only; track the ones they name; research-competitors does the rest. |
**Recorded as unknown, asked later by the playbook that needs it:**
| Fact | Asked by (internal; in the record write "when you …") |
|---|---|
| The goal number, deadline and budget | where-to-start (situation questions) or plan-campaign |
| Margins, most we can pay per customer | plan-campaign, run-paid-campaigns, email-lifecycle |
| What the agent may do alone, spend caps | run-paid-campaigns |
| This week's focus, next big date, live promotions | plan-campaign, email-lifecycle |
| Account access | whichever playbook needs the account |
| Search terms | improve-search-visibility |
| How results are counted | analyze-performance |
| Creators, partners, press relationships | work-with-creators |
| Story, origin and beliefs | make-content |
If the founder volunteers any of these, record them with provenance. Never refuse a fact; just don't ask for it.
## How to ask
- **Read first.** Documents the founder hands over are imported before any question (step 2). Never ask anything the site, the brain or an uploaded document answers.
- **Draft, then confirm.** "I have your customer as: women 25–40 with adult acne who've tried prescriptions and hated the dryness. Right, or closer to something else?" Never ask an open question when a draft is possible.
- **Group, don't interrogate.** One message per group (what we sell; who it's for; claims by risk; never-say; voice). Each item has a suggested answer, why it matters in a few words, and an easy "don't know". The basics take about 5–8 decisions and under 10 minutes; the whole run about 15–20 decisions and 20–30 minutes (SKILL.md time budget).
- **No jargon** the founder didn't use first: no ICP, CAC, LTV, USP.
- **Never ask:** for three adjectives; "what's your USP / target audience / brand personality?" as open questions; rating scales; passwords; anything already answered this session.
- **"Don't know" is an answer.** Record it as unknown with the assumption in use. It is never re-asked next session unless something changed.
## Contradictions
Surface every contradiction you find; never pick a side silently.
- **Precedence when the owner is unavailable:** what the owner says > the owner's documents > rules the owner made earlier (in learnings) > the brand's own site > the server's research. Inside the research itself, look for clashes too: a price on the site vs in an ad, an old tagline still live, a product the catalogue lists that the site no longer sells.
- **Show both sides** with source and date, the interim default, and the question: "Site footer (read 14 Oct): free shipping over $45. Your product page banner: over $35. Which is current? Until you say, nothing will mention a threshold."
- **Saved sources.** `brand_read` `sections: ["sources"]` lists the social accounts and review profiles research found, with `confirmed: false`. Show the brand's own ones (relationship `self`) and ask the owner to confirm they are really theirs. A wrong handle poisons the voice read and any later review import.
- Log each resolution in the brand record: who decided, when, in their words.
## Provenance for every fact
Every fact in the brand record carries:
- **Source:** told by owner / told by agency / approver via agency / owner's document / found on site / server research / agent note (unconfirmed) / inferred.
- **Date** captured, and **confirmed by** whom and when.
- **Visibility:** public / internal only / embargoed until a date / under NDA. Revenue, margin, budget and customer names default to internal.
- **Scope:** whole brand / one product / one audience / one market or location.
## The unknowns ledger
One line per gap:
- the gap (and its knowledge section);
- the job it blocks, and whether it blocks or only degrades that job;
- the assumption in use meanwhile, labelled **guess**;
- who was asked and when (or "not asked yet"), and when it will be asked, in plain words ("when you plan a campaign"): the owner reads the ledger, so no playbook names in it;
- what reopens it.
## Company-type add-ons
Ask these only when they apply **and** change a claim, an offer or a rule now; record the rest as unknown. Each is still draft-then-confirm.
- **DTC / e-commerce:** returns, shipping thresholds and zones, guarantee; standing discount rules and the deepest discount allowed; subscription, loyalty and referral programmes with their rewards and tiers (so no email promises the wrong reward); hero SKUs and never-advertise SKUs.
- **B2B SaaS:** which customers may be named (default none; written permission per name, with who granted it and when; design partners are a material connection) and which case studies are cleared; security and compliance certificates as expiring claims; competitor-naming policy; public pricing vs "talk to sales"; plan and tier names; roadmap items that are embargoed; analyst-quote rules.
- **Consumer app:** store ratings per platform and country (as expiring claims); trial and pricing terms per platform; kids or health-data flags.
- **Marketplace seller:** which claims are allowed on each marketplace (stricter than the own site); listing copy pasted or uploaded by the owner (the agent cannot read Amazon pages); authorised sellers and minimum advertised price policy.
- **Local or multi-location:** per-location name, address, phone, hours, services and offers; who controls local marketing (franchise or head office). The brain has no per-location shape today: record each location in the brand record and say so.
- **Services firm:** service packages and what each includes; team credentials and bios; client-naming default none; professional advertising rules for the profession.
- **Regulated category** (beauty, supplements, food, finance, alcohol, kids, health care, weight loss, CBD): load the rule pack from category-rules.md before drafting any claim; claims default to product facts only.
- **Several markets or languages:** tag claims, prices and rules by market.
Related skills
| Skill | Stars | Updated | Works with | Status | Install |
|---|---|---|---|---|---|
| rea-analysis-optimizationmorluto/reaOptimize REA providers and shared analysis workflows for agent context, latency, CPU, and memory while preserving correctness and Evidence. Use for analysis bottlenecks and cache or query design; CI/test-suite optimization belongs to rea-ci-tests-audit. | Stars: 472k | Updated | Claude CodeCodex | Verified | Install |
| rea-ci-tests-auditmorluto/reaAudit REA CI and test suites for correctness, fixture drift, false passes, and measured performance; use for suite or workflow investigations and efficiency changes, rather than ordinary feature testing or a single test failure. | Stars: 472k | Updated | Claude CodeCodex | Verified | Install |
| rea-tool-designmorluto/reaAudit, investigate, design, or change REA investigation tools, CLI/MCP contracts, provider capabilities, and Evidence semantics. | Stars: 472k | Updated | Claude CodeCodex | Verified | Install |
| ponytail-auditDietrichGebert/ponytailQuality audit of the whole repo: bugs, security, real load, missing tests, speed, and what to delete. Most important first. | Stars: 196k | Updated | Claude CodeCodex | Verified | Install |