clay-run/agent-plugins/onboard
Claude Code and agent skill by clay-run/agent-plugins
- community
Checking that the SKILL.md file is still in the repository.
Clay first-run onboarding — banner, what Clay can do, and starter tasks served by the Clay API. Run this right after a sign-in that just happened — from `clay login` directly, or invoked by the setup skill after its sign-in step. It checks server-side whether this user has been onboarded and stands down if so. Assumes the CLI is already signed in; for a signed-out user, run `setup` instead — it chains into this.
Install
npx vskill@latest install clay-run/agent-plugins/onboardSKILL.md
Clay onboarding
Give a brand-new user a fast first win. Whether this user has already been onboarded is tracked server-side, per user: clay login and clay whoami report it as onboarded, and fetching the starter-task options marks them as onboarded. There is nothing to record locally.
This skill runs signed-in. The setup skill is the entry point for anyone who isn't: it signs the user in and invokes this skill right afterward. If a command below fails with an auth error, run setup and stop — it will come back here when appropriate. If bare clay isn't on PATH this session (a fresh install — setup reports an absolute executable path for this case), use that resolved path in place of clay for every command in this skill.
The user's own request always wins. If sign-in was just a prerequisite for something concrete they asked for ("get this person's work email", a table query, a fix for a broken tool), do not run the onboarding menu now — finish their request; their own task is a better first win than a menu. After it's done, a one-line offer of the starter tasks is fine if it fits the moment. Since the options were never fetched, they stay un-onboarded server-side, so a future login can still make the offer cleanly. Run onboarding immediately only when there's no concrete task waiting: the user asked to get started, just installed the plugin, or setup itself was the whole request.
Keep your own prose short — the banner and the starter tasks are the show. The welcome (steps 2–3) goes on screen before you fetch anything; never repeat it within a session.
1. Check that onboarding applies
If the sign-in that brought you here already reported onboarded: false, skip this check and continue. Otherwise read the field yourself:
- `onboarded: false` — continue to step 2.
- `onboarded: true` — they've been onboarded before (possibly on another
machine). Stand down: skip the rest of this skill without ceremony. If the banner already went up (the Claude Code hook shows it at invocation), a one-line welcome back is enough — no pitch, no menu.
- The field is absent — an outdated CLI omits it too, so don't conclude the
server predates onboarding until you've ruled that out: compare clay --version against the plugin's minimum cli-min-version. If it's older, run the setup skill with that context — it installs a compatible independent CLI (its step 3) — then re-run clay whoami through that executable and read onboarded from there. If the versions match (or the CLI's whoami still omits the field), the server predates onboarding — stand down as in the true case.
2. Show the banner
On Claude Code, when this skill was invoked through the Skill tool and command -v jq succeeds, skip this step: a plugin hook displays the banner automatically at that invocation (it needs jq to do so) — never print it yourself; the step-3 pitch still opens your reply. In every other case — no jq, any other host, or following this file directly as a runbook (no Skill tool call, so no hook fired) — print a banner file from this skill's directory verbatim inside a fenced code block, before any other tool call (AskUserQuestion included). Pick the file by terminal width: run printf '%s\n' "${COLUMNS:-0}" and use banner.txt only when it prints a number of 115 or more; on anything smaller, non-numeric, or a failed probe, use banner-narrow.txt — it fits every host. Exception: a number from 1 to 39 means the pane is too narrow even for the narrow art — show no banner at all and go straight to step 3. Do not redraw, trim, or restyle the art. Reading the file is not showing it: the banner must appear in your reply to the user.
3. Say what Clay is
Position it as Clay's product and theirs to run ("Clay lets you…", "you can…"), not as your own abilities. Two or three sentences: Clay is a GTM data and automation platform — search hundreds of millions of people and companies, enrich them with verified emails, phone numbers, firmographics, and tech stacks, and automate the whole thing with functions and workflows.
Then show the breadth in a compact "you can ask me to…" list — four or five one-line bullets, phrased as plain asks, e.g.: find people or companies matching a profile, get verified emails for a list, see what technologies a company uses, query and export your existing Clay tables, build a workflow around a webhook, check your credit balance. No CLI command syntax and no skill names — the user talks to you, not to the CLI. If they ask for the full command surface, show them live clay --help output instead of reciting a remembered list.
4. Fetch and offer the starter tasks
This returns the current option set ({ data: [{ id, label, description }] }) and marks the user as onboarded server-side. Present exactly what it returns — the options are server-controlled and may differ from anything you remember. If the AskUserQuestion tool is available, use it (single-select) with each option's label and description, and don't add a free-form option of your own ("Type something", "Other") — the tool renders its own escape hatch. Without the tool, present a numbered list ending with one extra line: they can also just say what they'd like to do instead.
If the command is missing (command not found, or an unknown-command error naming onboard), the installed CLI predates onboarding. If the setup skill hasn't already run in this session, run it now with that context — the CLI on PATH is outdated, so it must install a compatible independent CLI (its step 3) even though sign-in checks pass — then retry clay onboard options once and continue normally if it works. Retry via the verified CLI's absolute path when bare clay still resolves to the old install (no restart needed for that). If setup already ran or the retry still fails, treat it like any other failure below.
For any other failure (network, server trouble), don't block and don't diagnose — no error-hunting, no environment debugging, no mention of what went wrong. Just ask the user what they'd like to do with Clay and help with that directly; the rest of this skill doesn't apply.
5. Record the selection and run the task
For an option from the list:
This records the pick and returns instructions — follow them as if the user had asked for that task directly. If the independent CLI's directory is not on PATH in this session, use the verified absolute executable path from setup for the selected task too.
If select fails, proceed anyway: treat the option's label as the user's request — but still get the user's explicit go-ahead before running anything that spends credits; the guidance a successful select returns requires that, and its absence doesn't waive it. Never block on the recording, diagnose it, retry it more than once, or mention it to the user.
If the user picks something outside the list (e.g. "Other"), skip select and just help with what they asked.
Security
- Security scanPassed, 100 of 100Pattern checks on the skill's files.
- Model reviewNot runA model review is scheduled for the next rescan.
Quality evaluation
- Task evaluationNot runNo evaluation yet. Run one in Skill Studio.
Source
- SKILL.md in the repositoryCheckingclay/skills/onboard/SKILL.md
Popularity
- Repository stars
- 143
- Repository forks
- 22
- 7-day trend score
- 0
Stars and forks are for the whole GitHub repository, not this skill alone. Not refreshed yet.
Works with
Any agent vskill installs to: 56 targets in all. These 39 have a page on this site.
- Claude Code
- Cursor
- GitHub Copilot
- Windsurf
- Codex
- Gemini CLI
- Cline
- Amp
- Roo Code
- Goose
- Continue
- Kilo Code
- Junie
- Trae
- OpenHands
- Qwen Code
- AI Agent
- Tabnine Chat
- Sourcegraph Cody
- Amazon Q
- Firebase Genkit
- Sweep
- Devin
- Phind
- Mentat
- Claude Desktop
- Aider
- GPT Pilot
- Plandex
- bolt.new
- v0
- Replit Agent
- Anse
- TypingMind
- BetterChatGPT
- LibreChat
- Warp AI
- Pieces
- Fig AI
Embed badge
[](https://verified-skill.com/skills/clay-run/agent-plugins/onboard)