proofly-automation

Overview of proofly-automation skill

What proofly-automation does

proofly-automation is a Claude skill for running Proofly operations through Composio’s Rube MCP server. Instead of hard-coding Proofly API calls, the skill instructs the agent to discover the latest available Proofly tools with RUBE_SEARCH_TOOLS, check the user’s Proofly connection, and then execute the selected workflow through Rube.

This makes the proofly-automation skill most useful when you want an AI agent to automate Proofly work but do not want to manually inspect changing tool schemas before every task.

Best-fit users and workflows

Use proofly-automation if you already work with Proofly and want Claude or another MCP-capable agent to help with repeatable Proofly tasks. It fits users who need workflow automation, operational execution, or tool-assisted actions rather than a plain text explanation of what to do.

The best-fit reader is someone who can connect MCP tools, authorize a Proofly account, and describe the target Proofly outcome clearly enough for the agent to select the right Rube tool.

What makes this skill different

The key differentiator is the “search tools first” pattern. The skill does not assume today’s Proofly tool schema will remain stable. It tells the agent to call RUBE_SEARCH_TOOLS before execution, then use the returned tool slugs, input schemas, execution plan, and pitfalls.

That matters for install decisions: this is not just a prompt template. It is a workflow guardrail for using live Composio/Rube tool discovery safely.

Important adoption requirements

Before installing, confirm that your client supports MCP and can connect to Rube at https://rube.app/mcp. You also need an active Proofly connection managed through RUBE_MANAGE_CONNECTIONS with toolkit proofly.

If you cannot use MCP tools, cannot authorize Proofly, or only need generic Proofly advice, this skill will not add much value over a normal prompt.

How to Use proofly-automation skill

proofly-automation install and setup path

Install the skill in a compatible skills environment with:

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

Then configure Rube MCP in your client by adding:

https://rube.app/mcp

After MCP is available, verify that RUBE_SEARCH_TOOLS responds. Next, call RUBE_MANAGE_CONNECTIONS for toolkit proofly. If the connection is not ACTIVE, follow the returned authorization link and re-check status before asking the agent to run Proofly actions.

Inputs the skill needs from you

For strong proofly-automation usage, give the agent more than a vague command. Include:

• the Proofly outcome you want
• the target object, record, campaign, project, or account context if applicable
• constraints such as “preview before making changes” or “do not delete anything”
• whether the task should only search, update, create, export, or verify
• any known field names, IDs, names, dates, or filters

A weak prompt is: “Use Proofly.”
A stronger prompt is: “Using proofly-automation, discover the current Proofly tools, verify my Proofly connection, then find the relevant records for [specific goal]. Show the planned tool call and required fields before executing any write action.”

Practical workflow for agent calls

A reliable proofly-automation guide should follow this sequence:

1. Call RUBE_SEARCH_TOOLS with a use case matching the exact Proofly task.
2. Reuse the returned session ID for follow-up discovery where possible.
3. Call RUBE_MANAGE_CONNECTIONS with toolkits: ["proofly"].
4. Stop if the connection is not ACTIVE.
5. Select the Proofly tool based on the returned schema, not memory.
6. Ask the user for missing required fields before execution.
7. Run the tool and summarize the result, including what changed.

This pattern reduces failed calls caused by stale assumptions about tool names or required parameters.

Repository files to read first

The upstream skill currently centers on one file: SKILL.md under composio-skills/proofly-automation. Read it before use because it contains the operational contract: prerequisites, Rube setup, tool discovery, connection checking, and the core workflow pattern.

There are no extra rules/, resources/, references/, or scripts in the current file tree, so the main decision is whether the SKILL.md workflow is enough for your Proofly automation needs.

proofly-automation skill FAQ

Is proofly-automation for Workflow Automation?

Yes. proofly-automation for Workflow Automation is the right framing when you want an agent to execute Proofly tasks through MCP tools, not simply draft instructions. Its value is in orchestrating discovery, connection validation, schema-aware execution, and result reporting.

Can I use it without Composio Rube MCP?

No. The skill explicitly requires Rube MCP and expects RUBE_SEARCH_TOOLS plus RUBE_MANAGE_CONNECTIONS to be available. Without those tools, the agent cannot follow the intended Proofly execution path.

How is this better than an ordinary prompt?

An ordinary prompt may invent API fields, rely on outdated assumptions, or skip authorization checks. The proofly-automation skill forces the agent to discover the current Proofly tool schemas first and confirm that the Proofly connection is active before running workflows.

That is especially useful when tool schemas, available actions, or required fields change over time.

When should I not install it?

Do not install it if you only need a one-time explanation of Proofly, cannot connect MCP servers, or need a fully custom Proofly integration with code-level control. Also avoid using it for sensitive write actions unless your prompt requires preview, confirmation, and clear rollback expectations.

How to Improve proofly-automation skill

Make proofly-automation prompts more specific

The most common failure mode is asking for a broad Proofly task without enough execution context. Improve results by giving the agent a concrete goal, the relevant Proofly scope, and a safety policy.

For example: “Search tools for updating Proofly records related to [business context]. If a write-capable tool is found, list required fields first and wait for confirmation before executing.”

This helps the agent choose the right discovered tool and avoid premature actions.

Add safety checks before write actions

For any create, update, delete, send, publish, or bulk operation, require a plan before execution. Ask the agent to show:

• the discovered tool slug
• required inputs
• optional inputs it plans to use
• target records or filters
• expected side effects

This is the simplest way to make proofly-automation safer without modifying the skill itself.

Iterate from discovery results

After the first RUBE_SEARCH_TOOLS call, use the returned schema to refine your prompt. If the schema exposes required fields you did not provide, answer with exact values instead of asking the agent to infer them.

Good iteration: “Use the tool you found, set status to X, filter by Y, and do not modify records outside Z.”
Poor iteration: “Just continue.”

Improve the skill for team use

If your team uses Proofly repeatedly, consider extending the local skill instructions with your preferred approval rules, naming conventions, allowed operations, and examples of common Proofly workflows. Keep those additions separate from tool schemas, because schemas should still come from RUBE_SEARCH_TOOLS at runtime.

A strong team version of proofly-automation should preserve the current discovery-first design while adding organization-specific guardrails.