nitrosend-setup
Skill3 starsChanged 7 months agoMITView on GitHub
Guided Nitrosend onboarding. Walks through remote MCP connection, Brand Kit identity setup, brand-subdomain preparation or customer-domain verification, and sender defaults. Also offers to configure proactive analytics (daily/weekly/monthly reports). Use when: "set up nitrosend", "configure email", "onboard", "get started with nitrosend", or when nitro_get_status shows incomplete onboarding.
In this repository6 other files it gives its agents.
- build-flowSkill
- compose-emailSkill
- import-contactsSkill
- nitrosend-analyticsSkill
- send-campaignSkill
- send-transactionalSkill
The file
Show its rules and text
We split this skill into the rules it gives an agent
22 rules. About 0 are common, 8 are distinctive, and 14 are very specific to this project.
≈0
Common · estimate
Many files like this have a rule of this kind.
≈8
Distinctive · estimate
This author's own choices. Worth a look.
14
Very specific
Tied to their project's names, files or tools. Probably skip.
- Common
- Distinctive
- Very specific
- Very specific: line 18
If the user is in Claude Desktop, Claude.ai, Claude Cowork, or this Nitrosend Claude plugin, connect Nitrosend through the remote MCP …
- Very specific: line 28
Use the CLI only when the user explicitly wants a terminal, CI, or headless agent-runner workflow outside Claude Desktop.
- Very specific: line 35
Run nitro_get_status to see what's already configured.
- Distinctive: line 39
If the MCP server is connected, OAuth is already working.
- Distinctive: line 41
If not connected, tell the user:
- Very specific: line 48
Verify the connection with nitro_get_status
- Distinctive: line 52
If the user needs a different Nitrosend account on an OAuth connection:
- Very specific: line 54
Run nitro_get_status and read the ids from available_accounts.items[*].id
- Very specific: line 55
Run nitro_select_account with the target account_id — the switch takes effect on the next tool call and lands on that account's default …
- Very specific: line 64
If the brand has no real business context, ask the user for their website URL and run nitro_set_brand_kit with url, the default synchronous …
- Distinctive: line 69
If they don't have a website, ask for:
- Very specific: line 78
Read brand.welcome_email in status before composing.
- Very specific: line 90
Read brand_subdomain from nitro_get_status.
- Very specific: line 98
If needed, poll nitro_get_status until brand_subdomain.ready is true.
- Very specific: line 99
Only if it is not selected, call nitro_manage_domains with operation: "select_brand_subdomain" and the user's chosen local_part (default …
- Distinctive: line 109
Do not claim that account creation or a draft consumes sending infrastructure.
- Distinctive: line 110
Do not describe a shared sender or an allowance for unverified sending.
- Very specific: line 114
Use nitro_configure_account to set:
- Distinctive: line 133
Use only the scheduling operation actually exposed by the host.
- Distinctive: line 143
If they do not establish an answer, search again or say what is not verified — never invent product behavior.
- Very specific: line 150
Create a first email with compose-email (/nitrosend:compose-email)
- Very specific: line 151
Add an audience with import-contacts (/nitrosend:import-contacts)
≈ “Common” is an estimate: at least 1 in 5 of the 350,495 skill files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.
Files
Nitrosend Setup
Guide the user through Nitrosend onboarding. Check what's already done and skip completed steps.
Connection Surface
If the user is in Claude Desktop, Claude.ai, Claude Cowork, or this Nitrosend
Claude plugin, connect Nitrosend through the remote MCP connector at
https://api.nitrosend.com/mcp. Do not recommend a terminal CLI or API-key
setup for standard Claude Desktop onboarding.
In any other agent (Antigravity, Cursor, Codex, Gemini CLI, VS Code and others), connect the same remote MCP server through that client's MCP settings with browser OAuth. The per-client guides are at https://docs.nitrosend.com/integrations/overview.
Use the CLI only when the user explicitly wants a terminal, CI, or headless
agent-runner workflow outside Claude Desktop. If a command sandbox cannot reach
api.nitrosend.com, treat that as sandbox egress, not as a reason to switch
Claude Desktop users to the CLI.
Step 1: Check Current State
Run nitro_get_status to see what's already configured. Report the current state briefly.
Step 2: Connection
If the MCP server is connected, OAuth is already working. Confirm this.
If not connected, tell the user:
- Open the MCP or connectors settings in their client
- Select the bundled
nitrosendserver if a Nitrosend plugin is installed (Claude or Antigravity). Otherwise addhttps://api.nitrosend.com/mcpusing the guide for that client at https://docs.nitrosend.com/integrations/overview - Complete the browser sign-in and approval flow
- Verify the connection with
nitro_get_status
Multiple Accounts
If the user needs a different Nitrosend account on an OAuth connection:
- Run
nitro_get_statusand read the ids fromavailable_accounts.items[*].id - Run
nitro_select_accountwith the targetaccount_id— the switch takes effect on the next tool call and lands on that account's default brand - Re-run
nitro_get_statusto confirm, thennitro_select_brandif a non-default brand is needed
API-key connections are pinned to one account and cannot switch.
Step 3: Brand Kit
If the brand has no real business context, ask the user for their website URL and run
nitro_set_brand_kit with url, the default synchronous mode, and a stable
idempotency_key. Let them override any scraped values. Reuse that key only
for an exact retry.
If they don't have a website, ask for:
- Company name
- A sentence about what the business offers and who it is for
Save fields.company_name and fields.company_description. Use the theme
schema's default brand colour unless one is supplied. Do not ask for a hex
colour or logo before understanding the business, and never invent "My Brand"
or a description. A name and colour alone are not enough to write from.
Read brand.welcome_email in status before composing. A queued, writing or
sending preview is already being prepared on the server; let it finish and
read its campaign. If written or sent, show the existing draft. Do not create
or send another first email just because the agent has connected. This owner
preview never sends to subscribers.
Send direct identity values in fields. Pass a public logo_url separately;
remote logo fetches also require a stable idempotency_key. For a local image,
use nitro_ingest first and pass its returned media_url or image_url.
Step 4: Sending Domain
Read brand_subdomain from nitro_get_status. If there is no ready selected
sender, offer these two explicit paths:
- Recommended quick start: call
nitro_manage_domainswithoperation: "prepare_brand_subdomain". This reserves the brand-matched<name>.nitrosend.netaddress and creates its local sending identity under Nitrosend's verified root. No per-brand DNS or DKIM work is needed. It sends no email and preserves an existing sender selection. - If needed, poll
nitro_get_statusuntilbrand_subdomain.readyis true. Only if it is not selected, callnitro_manage_domainswithoperation: "select_brand_subdomain"and the user's chosenlocal_part(defaulthello). Never infer selection from readiness. - Customer-domain alternative: use
nitro_manage_domainswithoperation: "add", show the returned DNS records, wait for the user to add them, then useoperation: "verify". It names any record still missing and is safe to repeat; once the records are in place, Nitrosend verifies the domain on its own, andverifyshows its progress.
Do not claim that account creation or a draft consumes sending infrastructure. Do not describe a shared sender or an allowance for unverified sending.
Step 5: Sender Defaults
Use nitro_configure_account to set:
from_name: Their preferred sender namefrom_email: Must match the explicitly selected Nitrosend brand-subdomain sender or a verified customer domainreply_to: Where replies should gotest_email_recipients: Up to 5 emails for test sends
Step 6: Proactive Analytics (Optional)
Offer proactive analytics only if the current host exposes a scheduled-task capability. If it does not, say that recurring reports are unavailable in this client and continue without inventing a task tool. When scheduling is available:
- Daily (weekday mornings): Quick summary of new subscribers and any sending activity from yesterday. Use
nitro_get_insightswithperiod: "7d"andnitro_queryfor recent contacts. - Weekly (Monday): Open/click rates, list growth, deliverability summary. Use
nitro_get_insightswithperiod: "7d"for trends. - Monthly (1st of month): Deep dive using
nitro_get_insightswithperiod: "30d", comparing with the account's prior periods and any benchmarks returned by the live tool.
Use only the scheduling operation actually exposed by the host. Suggested cron expressions:
- Daily:
"23 8 * * 1-5"(8:23am weekdays) - Weekly:
"47 8 * * 1"(8:47am Mondays) - Monthly:
"13 9 1 * *"(9:13am on the 1st)
Product Questions
For any Nitrosend product question not answered by the current context (how a
feature works, setup or verification steps, API/SDK/CLI usage, integrations),
call nitro_search_docs and treat the returned excerpts as the source of truth.
If they do not establish an answer, search again or say what is not verified —
never invent product behavior.
Completion
Summarize what was configured and suggest next steps by workflow name. In the Claude plugin, include the slash command:
- Create a first email with
compose-email(/nitrosend:compose-email) - Add an audience with
import-contacts(/nitrosend:import-contacts) - Build an automation with
build-flow(/nitrosend:build-flow)
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.

