agencyzoom-automation

Overview of agencyzoom-automation skill

What agencyzoom-automation is for

agencyzoom-automation is a Claude skill for running Agencyzoom CRM operations through Composio’s Rube MCP server. Its main value is not a fixed set of hardcoded actions; it teaches the agent to discover the current Agencyzoom tool schemas first, verify the connection, and then execute the right Rube MCP tool for the requested workflow.

Use this skill when you want an AI assistant to help with Agencyzoom tasks such as CRM record operations, pipeline updates, account follow-up workflows, or operational checks where the exact available action depends on the live Composio toolkit schema.

Best fit for CRM Operations teams

The agencyzoom-automation skill is best suited for CRM Operations, agency admin, RevOps, and automation users who already rely on Agencyzoom and want Claude to act through an authenticated MCP connection. It is especially useful when your task needs current tool discovery instead of guessing API fields from memory.

It is less useful if you only need written advice about Agencyzoom, do not have Rube MCP available, or cannot authorize an active Agencyzoom connection.

Key differentiator: search tools before action

The central pattern is: never assume the Agencyzoom schema. The skill instructs the agent to call RUBE_SEARCH_TOOLS before execution so it can retrieve available tool slugs, required inputs, field names, recommended plans, and pitfalls for the current toolkit version.

That matters because CRM automation fails quickly when an agent invents field names, uses stale schemas, or tries to execute before authentication is complete.

What to inspect before installing

The upstream skill is compact and currently centers on one file: SKILL.md. There are no extra scripts, resources, rules, or README files in the skill folder, so the install decision is mostly about whether you use Rube MCP and Agencyzoom.

Read SKILL.md first to confirm the required MCP server, connection workflow, and tool-discovery pattern match your environment.

How to Use agencyzoom-automation skill

agencyzoom-automation install context

If your client supports skill installation from GitHub, install the skill from the Composio skill collection, for example:

npx skills add ComposioHQ/awesome-claude-skills --skill agencyzoom-automation

Then configure Rube MCP in your AI client by adding:

https://rube.app/mcp

The skill expects Rube MCP tools to be available, especially RUBE_SEARCH_TOOLS and RUBE_MANAGE_CONNECTIONS. Before using it for real work, ask the agent to verify that RUBE_SEARCH_TOOLS responds and that the Agencyzoom toolkit connection is ACTIVE.

Required connection and discovery flow

A practical agencyzoom-automation usage flow should look like this:

1. Confirm Rube MCP is connected.
2. Use RUBE_MANAGE_CONNECTIONS with toolkit agencyzoom.
3. If the connection is not active, complete the returned authentication link.
4. Call RUBE_SEARCH_TOOLS for the specific Agencyzoom job.
5. Review the discovered tool schema and required fields.
6. Execute only after the tool and inputs are confirmed.

A good first instruction is:

Use the agencyzoom-automation skill. First verify Rube MCP access, then check that the Agencyzoom connection is active. Search for the current tool schema for updating Agencyzoom CRM records before taking any action.

Turn a rough goal into a complete prompt

Weak prompt:

Update my Agencyzoom leads.

Better prompt:

Use agencyzoom-automation for CRM Operations. I need to update Agencyzoom lead records that match these criteria: [criteria]. First call RUBE_SEARCH_TOOLS for the exact Agencyzoom operation and show me the discovered tool name, required fields, and any missing inputs. Do not execute until I confirm the mapping. When ready, update only the records in this list: [record IDs or exported table].

The stronger prompt works better because it supplies the task boundary, asks for schema discovery, requires a review step, and prevents broad CRM changes before confirmation.

Files and source path to read first

Start with:

• composio-skills/agencyzoom-automation/SKILL.md

That file contains the operational contract: prerequisites, setup, tool discovery, and the core workflow pattern. Because the skill has no helper scripts or reference folders, you should not expect prebuilt business logic for your agency. You will provide the workflow rules, record criteria, approval steps, and data mapping in your prompt.

agencyzoom-automation skill FAQ

Is agencyzoom-automation enough without Rube MCP?

No. The skill requires Rube MCP. If RUBE_SEARCH_TOOLS is unavailable, the agent cannot follow the skill’s main safety pattern. You also need an active Agencyzoom connection through RUBE_MANAGE_CONNECTIONS before the agent can perform authenticated operations.

How is this better than a normal Claude prompt?

A normal prompt may ask Claude to “use Agencyzoom,” but it does not enforce live tool discovery. The agencyzoom-automation skill is valuable because it directs the agent to search current Composio tool schemas before choosing an action. That reduces schema hallucination and helps the agent adapt to changes in available Agencyzoom tools.

Can beginners use this skill?

Yes, if they can configure MCP and complete the Agencyzoom authorization flow. Beginners should start with read-only or low-risk checks, ask the agent to display the discovered tool schema before execution, and require approval before changing CRM data.

When should I not use this skill?

Do not use it for bulk CRM changes without a tested record list, approval step, and rollback plan. Also avoid it when your task is mainly strategic advice, training content, or documentation; the skill is designed for live Agencyzoom operations through Composio, not general CRM consulting.

How to Improve agencyzoom-automation skill

Improve agencyzoom-automation prompts with constraints

The biggest output-quality improvement is giving the agent precise operational constraints. Include:

• The Agencyzoom object or workflow you want to affect
• Record IDs, search criteria, or a source table
• Fields to read or update
• Values to write and how to handle blanks
• Whether the agent may execute or must wait for approval
• Any records, statuses, or pipelines to exclude

This keeps agencyzoom-automation focused on the intended CRM operation instead of making broad assumptions.

Reduce failure modes before execution

Common failure modes include inactive connections, stale assumptions about field names, missing required fields, and overly broad update criteria. Prevent them by requiring the agent to show the discovered RUBE_SEARCH_TOOLS result summary before calling an execution tool.

A strong safety instruction is:

If the current Agencyzoom schema does not expose the required operation, stop and explain the gap. Do not substitute a different CRM action without approval.

Iterate after the first tool result

After the first discovery or execution result, ask the agent to reconcile what happened against your intended workflow. Useful follow-ups include:

• “Which required fields are still missing?”
• “Which records were skipped and why?”
• “Show the exact mapping before the next update.”
• “Run a smaller batch first.”
• “Summarize the completed changes for audit notes.”

These iterations are especially important for CRM Operations teams because the cost of a wrong update is higher than the cost of an extra confirmation step.

Add team-specific operating rules

The upstream skill does not include agency-specific policies. To make agencyzoom-automation more reliable, add your own rules in the prompt or local skill notes: approval thresholds, naming conventions, allowed pipelines, excluded statuses, audit-log format, and batch-size limits.

For recurring use, maintain a short internal playbook that pairs common Agencyzoom tasks with approved prompts and required review checkpoints.