seedling crm

Managing a small business shouldn't feel like a full-time job in itself. This skill connects your AI agent to Seedling CRM through a hosted MCP server, letting you manage customers, tasks, and billing without leaving your terminal or IDE.

What it does

• Task Orchestration Lists overdue items, triages boards, and updates statuses as you work.
• Meeting Intelligence Extracts action items from transcripts and converts them into structured tasks.
• Client Preparation Synthesizes customer history, open projects, and upcoming meetings into a single brief.
• Financial Operations Manages invoices, tracks expenses, and sends payment reminders to clients.
• Project Tracking Updates high-level project states and logs agent activity for team visibility.

How it works

1. Connect Add the Seedling MCP server to your client using your unique API bearer token.
2. Daily Brief Ask for a rundown of today's meetings and high-priority overdue tasks.
3. Active Management Update task statuses to in_progress or review in real-time as the agent completes work.
4. Billing Workflow Create and send invoices or record expenses by confirming details with the agent before execution.

Frameworks & tools

Built for the Model Context Protocol (MCP). Works with Claude Code, Claude Desktop, Cursor, and any host supporting HTTP MCP transport.

Why this beats prompting it yourself

Generic prompts cannot see your calendar or customer history. This skill provides direct, authenticated access to your CRM data, ensuring task updates and invoices are grounded in your actual business records rather than hallucinations.

Use cases

• Morning Standup Get a filtered brief of meetings and tasks that must move today.
• Call Preparation Pull a customer's full context and open promises before a Zoom call.
• Post-Meeting Triage Process raw notes into actionable tasks with clear descriptions and success criteria.
• Invoice Management Identify unpaid invoices and send reminders in seconds.

Known limitations

Risky writes may be held for manual approval by SafeNode policy. Performance is subject to rate limiting; large batch operations should be spaced out.