yelp-automation
by ComposioHQyelp-automation is a Claude skill for Yelp workflows through Composio Rube MCP. It guides agents to search current tools, verify the Yelp connection, and execute with returned schemas.
This skill scores 66/100, which makes it acceptable but limited for directory listing. Directory users get enough evidence to understand that it helps agents route Yelp-related requests through Composio's Rube MCP with connection checks and tool discovery, but they should expect a thin wrapper rather than a fully self-contained Yelp workflow library.
- Valid skill frontmatter clearly identifies the trigger domain as Yelp automation through Rube MCP and declares the required `rube` MCP dependency.
- Prerequisites and setup steps explain that Rube MCP must be connected, Yelp authorization must be ACTIVE, and `RUBE_SEARCH_TOOLS` should be called before workflows.
- The core workflow pattern gives agents a repeatable sequence: discover tools, check the Yelp connection, then execute with current schemas rather than guessing stale parameters.
- Actual Yelp actions and schemas are not documented in the repository; the skill relies on live `RUBE_SEARCH_TOOLS` discovery for current tool slugs, inputs, and pitfalls.
- No support files, scripts, README, or install command are included, so adoption depends on users already being comfortable configuring Rube MCP and Yelp connections.
Overview of yelp-automation skill
What yelp-automation is for
yelp-automation is a Claude skill for running Yelp-related workflows through Composio’s Rube MCP server. Instead of hard-coding one Yelp API shape, the skill tells the agent to discover the current Yelp tools first, verify the Yelp connection, and then execute the task using the schema returned by Rube.
Use it when you want an agent to handle Yelp operations inside a broader Workflow Automation setup, especially where tool availability or input schemas may change.
Best-fit users and jobs
The yelp-automation skill is best for users who already run Claude or another MCP-capable client and want to connect Yelp actions into repeatable workflows. Typical jobs include finding the right Yelp tool for a task, checking whether a Yelp connection is active, and safely executing a Yelp operation after schema discovery.
It is most useful for operators, automation builders, local business researchers, and agent workflow developers who prefer tool-mediated execution over manually browsing Yelp.
What makes this skill different
The main differentiator is the “search tools first” pattern. The skill does not assume static Yelp tool names or inputs. It requires RUBE_SEARCH_TOOLS before execution so the agent can use the latest tool slugs, schemas, execution plans, and pitfalls returned by Rube.
That makes yelp-automation more reliable than a generic prompt such as “use Yelp to find businesses,” because it gives the agent a connection-check and tool-discovery sequence before attempting the task.
Important adoption constraints
This is not a standalone Yelp scraper, browser automation script, or direct Yelp API wrapper. It requires Rube MCP and an active Yelp toolkit connection through Composio. The repository path contains only SKILL.md, so the skill is lightweight: you should read the source before installing, but do not expect helper scripts, examples, or bundled reference data.
How to Use yelp-automation skill
yelp-automation install and setup context
Install the skill from the Composio skills repository:
npx skills add ComposioHQ/awesome-claude-skills --skill yelp-automation
Then add Rube MCP to your client configuration using:
https://rube.app/mcp
Before asking for Yelp work, confirm that the MCP server exposes RUBE_SEARCH_TOOLS. The skill also expects you to use RUBE_MANAGE_CONNECTIONS with toolkit yelp; if the connection is not ACTIVE, follow the returned authorization link and re-check the status before running any Yelp workflow.
Inputs the skill needs
A weak request is: “Find restaurants on Yelp.” A stronger yelp-automation usage prompt gives the agent the task, location, filters, output format, and any limits:
Use yelp-automation to find highly rated Italian restaurants in Austin, Texas. First discover the current Yelp tools through Rube, confirm the Yelp connection is active, then return up to 10 results with name, rating, review count, price if available, Yelp URL, neighborhood, and a short reason each result matches. Do not invent missing fields.
For business lookup or review-oriented tasks, add known identifiers, business name, city, category, date range, ranking preference, and whether you need raw tool output or a cleaned table.
Practical workflow to follow
Start by reading composio-skills/yelp-automation/SKILL.md. There are no extra README.md, rules/, resources/, or scripts in this skill folder, so the source file is the main operating guide.
A good execution flow is:
- Ask the agent to call
RUBE_SEARCH_TOOLSwith your specific Yelp use case. - Reuse the returned session ID when possible.
- Ask it to call
RUBE_MANAGE_CONNECTIONSforyelp. - If active, execute the selected Rube tool using the exact schema from discovery.
- Ask for a concise result summary plus any raw IDs needed for follow-up steps.
This sequence matters because Rube may return different tool names or required fields than you expect.
Prompt pattern for better results
For reliable yelp-automation for Workflow Automation, write prompts that separate discovery, execution, and formatting:
Discover the available Yelp tools for searching local businesses near
94103. Check the Yelp connection. If active, run the appropriate tool for “coffee shops open now with at least 4 stars.” Return a markdown table with name, rating, review count, address, phone, Yelp URL, and any missing fields markedUnavailable. Explain which Rube tool was selected and why.
This improves output quality because the agent has enough context to pick the right tool, avoid schema guesses, and produce data your downstream workflow can parse.
yelp-automation skill FAQ
Is yelp-automation a Yelp API client?
No. yelp-automation is a skill that guides an MCP-capable agent to use Composio’s Yelp toolkit through Rube MCP. The actual tools, schemas, and authorization are handled through Rube, not through code shipped inside this skill.
When should I not use this skill?
Do not use it if you need an offline dataset, high-volume scraping, direct control over Yelp API credentials, or guaranteed fields independent of Rube’s current toolkit schema. It is also a poor fit if your client cannot connect to MCP servers or if you cannot authorize the Yelp connection.
How is it better than an ordinary prompt?
An ordinary prompt may ask the model to reason about Yelp from memory or browse inconsistently. The yelp-automation skill forces a tool-first workflow: discover current Yelp tools, check connection status, and execute with the returned schema. That reduces hallucinated tool calls and makes the workflow easier to repeat.
Is the yelp-automation skill beginner-friendly?
It is beginner-friendly if you already understand MCP basics. The setup is short, but the user must be comfortable checking tool availability, authorizing a connection, and asking the agent to use returned schemas. Beginners should start with a simple search task before chaining Yelp results into larger automations.
How to Improve yelp-automation skill
Improve yelp-automation prompts with clearer constraints
The biggest quality gain comes from precise task constraints. Include location, category, result count, ranking criteria, required fields, and how to handle missing values. For example, “top-rated dentists in Denver” is less useful than “return 15 dentists within Denver, sorted by rating then review count, with phone, address, Yelp URL, and Unavailable for missing fields.”
Avoid common failure modes
Common problems include skipping RUBE_SEARCH_TOOLS, assuming an old schema, trying to run Yelp actions before the connection is active, or asking for fields the selected tool does not return. Tell the agent explicitly: “Do not execute until tool discovery and connection check are complete,” and “Use only fields present in the tool response.”
Iterate after the first output
After the first run, improve the workflow by asking for deduplication, stricter filters, raw identifiers for follow-up calls, or a different output shape. If results look thin, broaden the radius, category, or price filters. If results are too broad, add neighborhood, open-now status, minimum review count, or business attributes.
Improvements maintainers could add
The yelp-automation skill would be stronger with a small examples section showing common Yelp workflows, sample prompts, expected RUBE_SEARCH_TOOLS usage, and troubleshooting notes for inactive connections. A short “when not to use” section and example output tables would also help users decide faster whether to install the skill.