welcome
Skill2 starsChanged 50 days agoMITView on GitHub
First-run onboarding — greet a new user, explain what Archetype Talk is and whose it is, list capabilities, and route to setup or the demo persona. Runs via the SessionStart hook on fresh installs or on /welcome.
- Reads credentials
8 other projects use a version of this.
Adopted by
Most recently taken up by
- samsam007b/claude-code-standards-kittheir own version, first seen Nov 16, 2026
- microsoft/amplifier-app-nanoclawtheir own version, first seen Oct 13, 2026
- TravisCao/cleanresearchtheir own version, first seen Oct 13, 2026
- shorin-nikita/ai-teamtheir own version, first seen Oct 11, 2026
- tymrtn/tomorrowkit-for-mindstheir own version, first seen Aug 30, 2026
- Q00/ouroborostheir own version, first seen Jul 31, 2026
Other versions8 projects use a version of this.
Word for word · 0
None found.
Edited · 0
None found.
Wrote their own version · 8
- Q00/ouroboros★ 6.4ksame name
- microsoft/amplifier-app-nanoclaw★ 5same name
- tymrtn/tomorrowkit-for-minds★ 4same name
- adraalabs/skills★ 2same name
- samsam007b/claude-code-standards-kit★ 2same name
- omarrns/claude-code-skills★ 1same name
- shorin-nikita/ai-team★ 1same name
- TravisCao/cleanresearch★ 1same name
How matches are found. Word for word: the exact same text. Edited: shares passages with this file (“99% same” means 99% of its five-word phrases also appear here; 10% or more counts). Their own version: similar meaning, or, weakest, the same name. We can’t tell which project had it first. Projects are counted among the 200 most-starred candidates, and mentions are links to the repository we found, up to 50 per site, so both counts are a floor.
In this repository31 other files it gives its agents.
- AGENTS.mdAGENTS.md
- CLAUDE.mdCLAUDE.md
- .github/copilot-instructions.mdCopilot instructions
- affinitySkill
- ai-personaSkill
- analytics-syncSkill
- backlogSkill
- catalogSkill
- cold-startSkill
- contradictionsSkill
- demo-dataSkill
- dovetail-syncSkill
- exportSkill
- extract-findingsSkill
- feature-panelSkill
- forget-participantSkill
- graph-lintSkill
- interview-guideSkill
- journey-mapSkill
- opportunity-treeSkill
- persona-avatarSkill
- persona-querySkill
- persona-talkSkill
- persona-voiceSkill
- persona-workshopSkill
- prdSkill
- prioritizeSkill
- researcherSkill
- retentionSkill
- source-syncSkill
- subject-accessSkill
The file
We split this skill into the rules it gives an agent
6 rules. 0 are common, 5 are distinctive, and 1 is very specific to this project.
0
Common
Most of the 7 edited versions we read keep these.
5
Distinctive
This author's own choices. Worth a look.
1
Very specific
Tied to their project's names, files or tools. Probably skip.
- Common
- Distinctive
- Very specific
- Very specific: line 14
Do this before you write a word: Read project_name: in Product Context.md and check .claude/cache/graph-index.md for its project: and …
- Distinctive: line 32
Keep it warm and short. Do not dump everything at once.
- Distinctive: line 36
Read README.md for the owner/author line and license.
- Distinctive: line 55
Ask the user (one clear question):
- Distinctive: line 60
Only what the project genuinely needs to continue:
- Distinctive: line 73
When onboarding is done, create the marker so it doesn't repeat:
“Common” is measured: kept by at least 6 in 10 of the 7 edited versions of this file we could read. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.
Files
welcome — First-run onboarding
Greet a new user right after they install/clone this repo, explain what it is, describe what you can do, and route them to either configuration or hands-on work. Runs automatically on first session (via the SessionStart hook, when .claude/onboarded.local is missing) or on demand with /welcome.
Adapt to the user's language. The repo ships in English by default, but detect the user's language from how they write and respond in it. Translate section labels and your prose; never translate verbatim participant quotes (they are data).
Step 0 — find out whether this repo is already somebody's project
Do this before you write a word. Read project_name: in Product Context.md
and check .claude/cache/graph-index.md for its project: and demo_entities:
lines. A repo that already carries a project is not a fresh install, and the
worst opening you can give is a tour of the demo to someone whose own research
is sitting right there.
- A project is named and real (non-
demo: true) entities exist → say what the project is in your first sentence, fromProduct Context.md. Offer to pick up the work; keep the capability list short and skip the demo entirely unless they ask. Never present a demo persona alongside their own. - Nothing but the demo → the flow below, as written.
The rest of this skill assumes the second case. In the first, use the parts that apply and drop the rest — an onboarding that ignores what is already in the folder is worse than no onboarding.
The onboarding flow — four steps, one at a time
Keep it warm and short. Do not dump everything at once.
Step 1 — What this is & whose it is
- This is Archetype Talk — a UX-research knowledge graph you can talk to. It turns real customer research into intention-based personas grounded in traceable evidence.
- Read
README.mdfor the owner/author line and license. State whose repo it is (author from README / LICENSE) so the user knows they're standing in someone's method, not a blank tool. - One-sentence shape:
Evidence → Signal → (Persona ↔ Archetype) → Idea, with optionalCompetitors/context.
Step 2 — What you can do here
Briefly list your capabilities (map to the skills):
- Talk to a persona grounded strictly in the research (
persona-talk) — incl. a ready DEMO persona:Personas/Emma.md. - Query a persona for likely reactions with confidence levels (
persona-query). - Build / deepen personas from research (
persona-workshop,ai-persona). - Start from absolute zero — no transcripts, no reports at all:
/cold-startinterviews the founder about the startup (value proposition, target group, pitch deck) and turns their beliefs into testable L1 assumptions + a research plan. Explicitly invoked only. - Turn transcripts into Signals & Evidence (
extract-findings). - Connect research sources over MCP — Dovetail (
dovetail-sync); voice memos & call recordings, Notion/Confluence, usability tools (source-sync); product analytics: Mixpanel / GA4 / Amplitude (analytics-sync). All read-only, with provenance. - Keep the graph healthy —
/graph-lint(dead links, orphans, stale data),/contradictions,/backlog. - Close the research loop —
/interview-guideturns backlog gaps into a discussion guide for real interviews. - PM toolkit —
/feature-panel(one feature × all personas, cited),/prioritize(RICE with Confidence derived from Levels),/prd(fully-cited Why),/opportunity-tree(OST from Correlations). - Deliver outward —
/journey-map(cited journey maps),/export(persona one-pagers, Ideas as tickets, stakeholder packs). - Give a persona an ElevenLabs voice so
/persona-talkreplies can be heard, not just read (persona-voice) — needs the user's own ElevenLabs API key. - Everything is graded by grounding Levels (L1–L5, one scale with persona maturity) and freshness (3-month rule), and every claim links back to a real source.
Step 3 — Ask which way to go
Ask the user (one clear question):
"Would you like to (a) go through a short setup so the repo fits your project, or (b) jump straight in — talk to the DEMO persona or add your own research?" Mention explicitly that they can try the DEMO persona right now with
/persona-talk Personas/Emma.md, no setup needed.
Step 4 — If they choose setup, ask the essentials (one at a time)
Only what the project genuinely needs to continue:
- Product under study — what product/service are the personas about? (updates
Product Context.md) - Research source — do they have transcripts to drop in
Inbox/, or a Dovetail workspace to connect via/dovetail-sync? (or neither yet → offer/cold-startfor a zero-data founding brief, or/persona-workshopto sketch an archetype) - Primary segment(s) — who are they researching first? (shapes the first persona/archetype)
3a. Competitor compare categories (only if they'll track competitors) — which 3–6 axes genuinely decide wins and losses in their market (pricing model, the contested capability, ownership, support, compliance…)? Write them to
compare_categories: [...]inProduct Context.mdfrontmatter — the app's Compare view shows exactly these. Steer them away from vanity metrics (funding, headcount, follower counts): if it doesn't change what their users choose, it's not a category. - Keep, separate, or clear the example data? — the Spotify example set is illustrative; every file carries
demo: truein frontmatter (a Demo badge in the app, ⚠️ flagged bypersona-talk). Three options via/demo-data: keep in place (badges mark it), separate it underDemo/(recommended once real data starts flowing), or delete it. No pressure to decide now —persona-talkwill ask once after their first demo session anyway. Real content they add should simply omit thedemo:field. - Language — confirm the language they want to work in (you adapt; repo stays English by default).
- Voice (optional) — want personas to have an ElevenLabs voice? Needs their own API key in
.env(ELEVENLABS_API_KEY); if they don't have one yet, skip this for now — it's revisitable anytime viapersona-voice, not a one-time choice.
Then propose the concrete next action (connect Dovetail / process Inbox / start a workshop) and, if useful, remind them the DEMO persona is available.
Finish
When onboarding is done, create the marker so it doesn't repeat:
touch .claude/onboarded.local(If the user only ran /welcome manually and was already set up, don't recreate the marker logic — just help.)
Rules
- Never fabricate the author/product — read
README.md/LICENSEfor the owner; if the product isn't set yet, say so and offer to set it inProduct Context.md. - Don't auto-run heavy work. Onboarding never bulk-analyzes files on its own — it points to
/extract-findingsand asks first. - One question at a time in Step 4; summarize what you heard before moving on.
- Never offer the demo to a repo that already has a project. Step 0 decides
this, and
demo: trueis the only thing separating the example set from the user's own files (hard rule 9 in CLAUDE.md) — so an unfiltered list of personas is always wrong here.
Does it work?
Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.
No reviews yet
People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.
- It worked
- Partly
- It didn't
Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.
Your trail
Files you open will collect here, so you can find them again.
Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.
Now do this for your own.
Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.

