Menü
Share a Clay workflow you built as a community template - use when a user wants to publish a table/spec they built as a reusable template, or asks how to contribute to community/templates.
Mit Codex oder Claude installieren Kopieren Sie diesen Prompt, fügen Sie ihn in Codex, Claude oder einen anderen Assistant ein und lassen Sie die Skill-Seite prüfen und installieren.
Basierend auf der SOC-Berufsklassifikation
| name | clay-share-workflow |
| description | Share a Clay workflow you built as a community template - use when a user wants to publish a table/spec they built as a reusable template, or asks how to contribute to community/templates. |
Sharing is a guided wizard, not a copy-paste. It rebuilds a clean template from your workflow's structure and logic - it never copies raw column config, which is where API keys and client data hide in a real Clay table.
node scripts/contribute.js
This is also wired into package.json, so it's runnable as npm run share.
Two modes:
node scripts/contribute.js # guided from scratch, Q&A in the terminal
node scripts/contribute.js --from my-spec.yaml # rebuild from an export-spec file
Use --from if you already have a spec from node clay-v2.js export-spec ...
(/clay-build-table) - the wizard reconstructs steps and input variables from it
automatically instead of asking you to describe each column by hand.
The wizard never touches your live table config directly. For each field it sees, it
classifies the field kind (action / formula / use-ai / http-placeholder) and
rebuilds only the parts safe to share:
{{placeholder}}.use-ai prompts get real IDs, emails, and URLs inside the prompt text swapped for
{{table_reference}} / {{email}} / {{url}} - but the wizard prints a note asking
you to reread each prompt anyway, since prompts remember company names in ways no
regex catches.http-placeholder steps are stripped to nothing but a note: configure your own
endpoint, keep credentials in env vars, never in the column.This is why the output is a template, structurally incapable of holding the things
that leak (see community/schemas/template.schema.json - there's no field a raw
export or a long token can even go into).
Before anything is written to disk, the wizard prints the exact
community/templates/<id>/template.yaml (and, if you opted into a contributor
profile, community/contributors/<handle>/profile.yaml) it's about to create, and
asks:
Read the preview like a stranger would. Company names in prompts? Real IDs? Anything you would not put on a billboard?
Read it that carefully before answering y. If anything looks off, answer n (or
just don't type y) - nothing is written, and you can delete the template's id
folder and start over.
Optional nicer preview: if npx -y lavish-axi is available in your environment,
you can render this preview as an HTML page for easier review instead of reading raw
terminal YAML - plain terminal preview is the default and always works, so treat this
as a convenience, not a requirement.
The wizard asks once: "Create a public profile? [Y/n]" - a public page with your name,
company, and LinkedIn that your templates link to and that ranks you on the front-page
leaderboard by vote totals. It is entirely optional: answer n and your template ships
with author: anonymous. If you opt in, only a GitHub handle and first name are
required; company, role, last name, and LinkedIn are all optional.
community/templates/<id>/template.yaml (and the profile file,
if any).node scripts/scan-repo.js
node scripts/validate-community.js
id folder to start
over) - nothing partial gets left in a broken state for you to accidentally commit.git checkout -b share/<id>
git add community/templates/<id>/ community/contributors/<handle>/
git commit -m "template: <title>"
git push and open a pull request
--from only works against a spec file (export-spec output), not a live table ID
/clay-build-table, then share.Set up auth, session, and profile config for clay-cli before running any command - use when a Clay command fails with "session expired", when starting fresh in a new checkout, or when configuring workspace/folder/workbook IDs for the first time.
Discover Clay actions and manage the local action catalog - use when searching for an available action/integration, checking catalog coverage against the strict proof registry, classifying action safety, or generating action templates.
Extract Clay table/workbook data to local files - use when you need a full-fidelity workbook export, a template-shaped snapshot of a table, or raw paginated row dumps for offline processing.
Run a full Clay playbook end to end - intake, plan, prepare, sample, verify, scale - use when a plain-language request should be routed to one of the 12 packaged playbooks and operated through to a safe scale decision, not just a single ad-hoc command.
Build Clay list-building sources - company-to-people waterfalls, source preview/import, and webhook intake - use when adding a new list-building source to a table or converting a companies table into a people search.
Build and modify Clay tables, views, fields, and records with clay-v2.js - use when creating a new table/workbook, adding or editing columns (text/formula/http-api/use-ai/action), managing views, or writing declarative table specs.