bubble-automation
by ComposioHQbubble-automation helps agents automate Bubble operations through Composio's Bubble toolkit via Rube MCP, using schema-first tool discovery, connection checks, and safe execution patterns.
This skill scores 66/100, which makes it an acceptable but limited directory listing. It gives agents enough trigger and setup guidance to work with Bubble through Rube MCP, but directory users should understand that it is a lightweight connector-oriented skill rather than a rich library of Bubble automation workflows.
- Valid frontmatter clearly identifies the skill as Bubble automation through Rube MCP and declares the required MCP dependency.
- Prerequisites and setup steps explain that Rube MCP must be connected, the Bubble toolkit connection must be ACTIVE, and RUBE_SEARCH_TOOLS should be called first.
- The skill gives a repeatable discovery-first pattern using RUBE_SEARCH_TOOLS and RUBE_MANAGE_CONNECTIONS, reducing guesswork compared with a generic Bubble prompt.
- No support files, scripts, references, or README are included beyond SKILL.md, so adoption depends entirely on the brief inline instructions.
- Workflow guidance is mostly generic tool discovery through Rube MCP; users looking for concrete Bubble task recipes or edge-case handling may find it thin.
Overview of bubble-automation skill
What bubble-automation does
bubble-automation is a Claude skill for automating Bubble app operations through Composio’s Bubble toolkit via Rube MCP. Instead of guessing Bubble API calls or relying on stale tool names, the skill’s central instruction is to discover the current Rube tools first, verify the Bubble connection, then execute the workflow using the returned schemas.
Best fit for Bubble and workflow automation users
This skill is best for builders, operations teams, and AI agents that need repeatable Bubble actions inside a broader Workflow Automation process: checking app data, creating or updating records, coordinating Bubble tasks with other SaaS tools, or turning a natural-language Bubble request into a tool-driven plan. It is most useful when your environment already supports MCP and you want the assistant to operate through authenticated Composio tools rather than only draft instructions.
Key differentiator: schema-first execution
The main value of the bubble-automation skill is not a large file set; the repository currently centers on SKILL.md. Its value is the workflow discipline it enforces: call RUBE_SEARCH_TOOLS before execution, use the current tool slugs and input schemas, manage the Bubble connection with RUBE_MANAGE_CONNECTIONS, and avoid hard-coded assumptions. That matters because MCP tool schemas and available Bubble operations can change.
Important adoption requirement
Do not install this expecting standalone Bubble automation without Rube MCP. The skill requires the rube MCP server and an active Bubble connection. If your AI client cannot call MCP tools, or if you cannot authenticate Bubble through Composio/Rube, this skill will not complete real actions.
How to Use bubble-automation skill
bubble-automation install and setup path
Install the skill in a compatible skills environment, for example:
npx skills add ComposioHQ/awesome-claude-skills --skill bubble-automation
Then add Rube MCP to your client configuration using the endpoint documented in the skill: https://rube.app/mcp. After the MCP server is available, verify that RUBE_SEARCH_TOOLS responds. Next, call RUBE_MANAGE_CONNECTIONS with toolkit bubble; if the connection is not ACTIVE, follow the returned authorization link and re-check status before asking the agent to modify anything in Bubble.
Inputs the skill needs from you
For strong bubble-automation usage, give the assistant a specific Bubble outcome, the target app or environment, object/data type names, field names, filters, and the acceptable level of change. A weak request is: “Update Bubble records.” A stronger request is: “Use the Bubble connection to find all Order records where status is pending_review and created_date is before 2025-01-01, then prepare a plan before changing status to archived. Search tools first and show the schema you plan to use.”
That extra detail reduces two common risks: selecting the wrong Bubble operation and running a broad mutation without enough constraints.
Practical workflow for safe execution
Start each task by asking the agent to run tool discovery for the specific use case, not just “Bubble operations.” The skill’s recommended pattern is:
- Search tools with
RUBE_SEARCH_TOOLS. - Inspect returned tool slugs, required fields, and warnings.
- Check the Bubble connection with
RUBE_MANAGE_CONNECTIONS. - Draft an execution plan using the current schema.
- Run read-only or narrow-scope actions first when possible.
- Execute write actions only after confirming the target records and fields.
For production apps, add an explicit confirmation gate: “Do not create, update, or delete records until you summarize the exact tool call and affected objects.”
Repository files to read first
Read composio-skills/bubble-automation/SKILL.md first; it contains the operational contract, prerequisites, setup notes, and example Rube calls. There are no extra rules/, resources/, references/, or scripts in the current file tree, so the skill is intentionally lightweight. If you need deeper Bubble toolkit coverage, follow the upstream toolkit documentation linked from the skill: https://composio.dev/toolkits/bubble.
bubble-automation skill FAQ
Is bubble-automation enough without Rube MCP?
No. The bubble-automation skill is an instruction layer for using Bubble through Rube MCP and Composio. It does not bundle its own Bubble API client, helper scripts, or offline executor. If RUBE_SEARCH_TOOLS and RUBE_MANAGE_CONNECTIONS are unavailable in your client, the assistant can only discuss the workflow, not perform it.
How is this better than a normal Bubble prompt?
A normal prompt may produce generic Bubble advice or outdated API assumptions. This skill tells the agent to discover live tool schemas before acting, which is important for reliable Workflow Automation. The practical benefit is fewer malformed tool calls and less guessing about required fields, connection state, or available Bubble operations.
Is the bubble-automation skill beginner-friendly?
It is beginner-friendly for Bubble users who already have an MCP-capable AI client and can complete an OAuth-style connection flow. It is less friendly if you are new to MCP, because the key setup step is outside Bubble itself: connecting Rube MCP, verifying tool availability, and activating the Bubble toolkit connection.
When should I not use this skill?
Do not use it for visual Bubble editor design, plugin development, complex database migrations without review, or broad production updates where you cannot define exact filters. It is also a poor fit when you need deterministic CI/CD-style deployment scripts; this skill is better for agent-mediated Bubble operations where tool discovery and human confirmation are acceptable.
How to Improve bubble-automation skill
Improve bubble-automation results with precise prompts
The highest-impact improvement is better task framing. Include the Bubble data type, operation, field names, filter logic, environment, and whether the agent should only inspect or also change data. For example: “Search Rube tools for Bubble record lookup. Find Customer records with plan = trial and last_login older than 60 days. Return a count first; do not update records until I approve.” This gives the skill a safer path from discovery to execution.
Avoid common failure modes
The most common failure is skipping RUBE_SEARCH_TOOLS and assuming a tool schema. The second is trying to run actions before the Bubble connection is ACTIVE. The third is asking for a large update without a verification step. Prevent these by making discovery, connection check, and confirmation part of every prompt. If an output seems too confident without showing tool discovery, ask the agent to restart from schema discovery.
Iterate after the first tool response
After the first Rube response, refine the plan using the actual returned schema. Ask the agent to identify required fields, optional fields, rate or pagination concerns, and any destructive operation warnings. For multi-step Workflow Automation, have it execute one narrow read action, summarize results, then proceed to writes only when the match set is correct.
Improvement ideas for maintainers and power users
The current skill is concise and usable, but it would become stronger with examples for common Bubble tasks such as listing records, creating records, updating fields, and checking connection status. A small troubleshooting section for inactive connections, missing tools, and schema mismatches would also improve adoption. If you fork or extend bubble-automation, keep the schema-first rule central rather than adding hard-coded tool calls that may become stale.