instacart-automation
by ComposioHQinstacart-automation helps agents run Instacart workflows through Composio Rube MCP by discovering current tool schemas, checking connection status, and requiring confirmation before sensitive actions.
This skill scores 68/100, which makes it acceptable but limited for directory listing. Directory users get a clear MCP-based entry point for Instacart automation and enough setup guidance to decide whether it fits, but the repository evidence is thin beyond SKILL.md and relies heavily on dynamic Rube tool discovery rather than packaged workflows or examples.
- Clear scope and trigger: it is specifically for automating Instacart tasks through Composio's Instacart toolkit via Rube MCP.
- Includes concrete prerequisites and setup checks, including RUBE_SEARCH_TOOLS availability, an active Instacart connection, and connection management before workflows.
- Emphasizes live tool discovery first, which helps agents avoid stale schemas and select current Instacart tool slugs and inputs.
- No support files, scripts, reference examples, or install command are included; adoption depends on users already knowing how to configure and operate Rube MCP.
- The workflow guidance is mostly discovery-oriented rather than task-specific Instacart automation, so agents may still need to infer exact actions from live tool schemas.
Overview of instacart-automation skill
What instacart-automation is for
instacart-automation is a Claude skill for automating Instacart-related workflows through Composio’s Rube MCP server. Its main value is not a fixed list of grocery actions; it teaches the agent to discover the current Instacart toolkit schema first, verify the user’s connection, then execute available tools with less guessing.
Best-fit users and workflows
This instacart-automation skill fits users who already work with MCP-enabled AI clients and want an agent to help with Instacart operations such as searching available toolkit actions, preparing structured tool calls, checking connection state, and running supported workflows. It is most useful when you want the agent to adapt to live Rube tool schemas instead of relying on stale hard-coded parameters.
Key differentiator: schema-first execution
The important design choice is mandatory tool discovery. The skill instructs the agent to call RUBE_SEARCH_TOOLS before attempting any Instacart action, because Composio tool names, inputs, execution plans, and pitfalls may change. This makes the skill stronger than an ordinary “use Instacart” prompt when tool reliability matters.
Adoption requirements to check first
Before installing, confirm you can use Rube MCP in your client and that an Instacart connection can be activated through RUBE_MANAGE_CONNECTIONS. The repository path contains only SKILL.md, so the skill is lightweight; there are no helper scripts, bundled references, or extra rule files to depend on.
How to Use instacart-automation skill
instacart-automation install context
Install the skill from the Composio skills repository, then add Rube MCP to your AI client:
npx skills add ComposioHQ/awesome-claude-skills --skill instacart-automation
Configure the MCP server endpoint as:
https://rube.app/mcp
The skill itself does not require a local code build, but it does require the MCP tools to be reachable. In practice, installation is only complete when RUBE_SEARCH_TOOLS responds in your client.
Connection setup before first usage
For reliable instacart-automation usage, ask the agent to verify the connection before running a task:
- Call
RUBE_SEARCH_TOOLSfor the specific Instacart use case. - Call
RUBE_MANAGE_CONNECTIONSwith toolkitinstacart. - If the connection is not
ACTIVE, follow the returned authentication link. - Re-check status before executing any Instacart tool.
This avoids the common failure mode where the agent drafts a plausible action but cannot run it because the Instacart toolkit has not been authorized.
Turn a rough goal into a complete prompt
A weak prompt is: “Use Instacart to add groceries.”
A stronger prompt for the instacart-automation guide is:
“Use the instacart-automation skill. First call RUBE_SEARCH_TOOLS for the current Instacart tools and schemas. Check whether my Instacart connection is active. I want to prepare a grocery order for delivery, prioritizing store availability, substitutions, and a final review before checkout. Do not execute irreversible actions until you summarize the planned tool calls and ask for confirmation.”
This works better because it states the workflow, asks for schema discovery, names decision constraints, and separates preparation from final execution.
Repository files to read first
Start with composio-skills/instacart-automation/SKILL.md. It contains the full operational pattern: prerequisites, setup, tool discovery, connection checking, execution, and confirmation. There are no README.md, scripts/, resources/, or rules/ folders in this skill, so do not waste time searching for hidden implementation files.
instacart-automation skill FAQ
Is instacart-automation beginner-friendly?
It is beginner-friendly for users who already understand MCP tool use, but not for users expecting a one-click Instacart bot. The skill assumes your AI client can access Rube MCP tools and that you can complete an Instacart authentication flow when prompted.
How is this better than a normal prompt?
A normal prompt may invent tool names or parameters. The instacart-automation skill explicitly requires RUBE_SEARCH_TOOLS first, so the agent should use current Composio schemas and available execution plans. That matters when APIs, fields, or supported actions change.
What should not be automated blindly?
Do not let the agent perform purchase-like, payment-related, account-changing, or irreversible actions without a review checkpoint. Use the skill to discover, prepare, compare, and draft actions; require explicit confirmation before checkout, order modification, cancellation, or address/payment-sensitive steps.
Does it work outside the Composio/Rube ecosystem?
Not directly. The skill is designed for Rube MCP and Composio’s Instacart toolkit. If your environment does not expose RUBE_SEARCH_TOOLS and connection-management tools, you can still reuse the workflow pattern, but the installed skill will not execute as intended.
How to Improve instacart-automation skill
Improve instacart-automation prompts with constraints
The skill performs best when you provide operational constraints up front: preferred store, delivery window, budget, dietary restrictions, substitution rules, quantity limits, and whether the agent may take action or only prepare a plan. These details reduce back-and-forth and help the agent choose the right discovered tools.
Avoid common failure modes
The most common issues are skipping tool discovery, using stale parameter names, assuming the Instacart connection is active, or treating a reversible planning step and an irreversible checkout step the same way. Reinforce the sequence: search tools, inspect schema, check connection, plan calls, ask for confirmation, then execute.
Iterate after the first tool result
After the first result, do not simply say “continue.” Ask the agent to compare returned data against your goal: unavailable items, substitutions, price changes, delivery constraints, or missing required fields. This turns tool output into a controlled workflow instead of a chain of blind API calls.
Extend the skill responsibly
If you customize instacart-automation, add examples that preserve its schema-first behavior. Good additions include confirmation rules, sample prompts for common shopping workflows, and safety notes for checkout-like actions. Avoid hard-coding current tool schemas into the skill; the repository’s core advantage is discovering them live through Rube MCP.