foursquare-automation
by ComposioHQfoursquare-automation helps agents automate Foursquare workflows through Composio Rube MCP by discovering current tools, checking active Foursquare connections, and using returned schemas before execution.
This skill scores 66/100, which makes it acceptable but limited for directory listing. Directory users can understand that it is meant to automate Foursquare through Composio/Rube MCP and it gives enough setup and discovery guidance for an agent to begin safely, but it offers little Foursquare-specific workflow substance or install-decision detail beyond generic tool discovery instructions.
- Valid skill metadata with a clear MCP requirement: it specifies Rube MCP and an active Foursquare connection as prerequisites.
- Provides an actionable setup path: add the Rube MCP endpoint, verify RUBE_SEARCH_TOOLS, manage the Foursquare connection, and confirm ACTIVE status before use.
- Strong trigger discipline: it repeatedly instructs agents to call RUBE_SEARCH_TOOLS first to retrieve current Foursquare tool schemas before execution.
- No support files, scripts, references, or README are present beyond SKILL.md, so adoption depends entirely on the brief in-file guidance.
- The content appears mostly as a generic Rube MCP/Foursquare toolkit wrapper; evidence shows low practical specificity and no concrete task examples beyond the discovery/check-connection workflow pattern.
Overview of foursquare-automation skill
What foursquare-automation does
The foursquare-automation skill helps an AI agent automate Foursquare actions through Composio’s Rube MCP server. It is built for workflows where the agent must first discover the current Foursquare tool schemas, verify an active Foursquare connection, and then call the right Rube tool rather than guessing parameters from memory.
Best-fit users and workflows
This skill is useful if you are using Claude or another MCP-capable agent to run Foursquare-related operations inside a broader workflow automation setup. Good fits include teams that want an agent to manage venue, location, discovery, or account-connected Foursquare tasks through Composio’s Foursquare toolkit, especially when tool schemas may change and need to be checked at runtime.
Key differentiator: tool discovery first
The main value of the foursquare-automation skill is not a static list of Foursquare commands. Its core instruction is to call RUBE_SEARCH_TOOLS before execution so the agent retrieves current tool slugs, schemas, execution plans, and pitfalls. That makes it safer than a generic prompt that assumes old API fields or invents unsupported parameters.
What to know before installing
The repository path contains only SKILL.md, so adoption depends on whether your client already supports MCP and whether Rube is configured. The skill requires the rube MCP server and an active Foursquare connection via RUBE_MANAGE_CONNECTIONS with toolkit foursquare. If you cannot add an MCP server or complete Foursquare authorization, this skill will not execute real actions.
How to Use foursquare-automation skill
foursquare-automation install context
Install from the skill repository with your skill manager, for example:
npx skills add ComposioHQ/awesome-claude-skills --skill foursquare-automation
Then add Rube MCP to your client configuration using:
https://rube.app/mcp
After installation, verify that the agent can access RUBE_SEARCH_TOOLS. The upstream skill states that no API keys are needed for the MCP endpoint, but Foursquare actions still require an active toolkit connection. Use RUBE_MANAGE_CONNECTIONS with toolkit foursquare, follow the returned auth link if needed, and confirm the connection status is ACTIVE.
Inputs the skill needs from you
For strong foursquare-automation usage, give the agent the actual Foursquare task, the account or venue context, the intended outcome, and any safety limits. A weak prompt is: “Use Foursquare.” A better prompt is: “Using the foursquare-automation skill, find the current Rube tools for Foursquare, confirm my Foursquare connection is active, then identify the tool and required fields for updating venue information. Do not execute changes until you show me the proposed parameters.”
This works better because the agent knows to discover tools first, check authorization, separate planning from execution, and avoid irreversible updates without confirmation.
Recommended workflow
Start with a planning pass:
- Ask the agent to read
composio-skills/foursquare-automation/SKILL.md. - Have it call
RUBE_SEARCH_TOOLSwith a specific use case, not a vague “Foursquare operations” query. - Have it inspect the returned schemas and recommended execution plan.
- Confirm the Foursquare connection through
RUBE_MANAGE_CONNECTIONS. - Execute only after the required fields, account context, and expected result are clear.
For example, use: “Search Rube tools for the use case ‘retrieve Foursquare venue details by location name and city,’ summarize the available tool schemas, then ask me for any missing required fields before calling a tool.”
Repository files to read first
Read SKILL.md first because it contains all available implementation guidance for this skill: prerequisites, setup, tool discovery, and the core workflow pattern. There are no separate README.md, resources/, rules/, or helper scripts in this skill folder, so do not expect bundled examples or custom validation code. The practical value comes from the runtime Rube tool discovery step.
foursquare-automation skill FAQ
Is foursquare-automation enough without Rube MCP?
No. The foursquare-automation skill is a routing and workflow instruction layer for Rube MCP. It does not include a standalone Foursquare API client, local scripts, or built-in credentials. You need an MCP-capable client, the Rube MCP endpoint, and an active Foursquare connection.
How is this better than an ordinary prompt?
A normal prompt may ask the model to “use Foursquare,” but it may guess tool names or outdated fields. This skill explicitly tells the agent to search tools first with RUBE_SEARCH_TOOLS, use the returned schemas, and verify the connection before executing. That reduces parameter errors and makes the workflow more resilient to toolkit changes.
Is this suitable for beginners?
Yes, if you are comfortable adding an MCP server and following an authorization link. It is not beginner-friendly as a standalone Foursquare tutorial. The skill assumes the agent can call Rube tools and that you understand the difference between discovering a tool, checking a connection, and executing an action.
When should I not use this skill?
Do not use foursquare-automation if you only need static Foursquare documentation, if your environment cannot run MCP tools, or if you require offline automation. It is also a poor fit for workflows where you cannot authorize a Foursquare account or where every action must be implemented as auditable local code rather than mediated through Rube.
How to Improve foursquare-automation skill
Make prompts specific to the Foursquare job
The best way to improve foursquare-automation results is to describe the exact job-to-be-done. Include whether you want to search, retrieve, create, update, or analyze Foursquare data; the venue, place, or location context; and whether the agent should execute or only prepare a plan. Specific use cases produce better RUBE_SEARCH_TOOLS results and cleaner schemas.
Add execution safeguards
For write actions, ask the agent to pause before execution and show the selected tool slug, required fields, optional fields, inferred values, and risks. A strong instruction is: “Do not call the final mutation tool until I approve the parameter object.” This is especially important for venue or account-connected changes where an incorrect field can affect live Foursquare data.
Handle common failure modes
The most common blockers are missing Rube MCP access, inactive Foursquare authorization, vague tool-search queries, and schema mismatch from skipped discovery. If a call fails, do not retry blindly. Ask the agent to rerun RUBE_SEARCH_TOOLS for the precise task, recheck RUBE_MANAGE_CONNECTIONS, compare the failed payload with the current schema, and then produce a corrected payload.
Iterate after the first output
After the first plan or tool call, refine with concrete feedback: “Use the returned schema only,” “ask for missing required fields,” “show the execution plan before acting,” or “limit results to this city/account/context.” The foursquare-automation skill works best when each step narrows uncertainty before the agent performs an external Foursquare action.