---
name: cars-jobs
description: "Step-by-step job recipes for the Car Check MCP server: Decode a VIN; Get a vehicle report; List recalls. Use when the user asks things like \"What car is VIN 1HGCM82633A004353?\" or \"Give me a report on a 2016 Honda Civic.\""
license: MIT
---

# Car Check jobs

Each section below walks through one job with the tools of the `cars` MCP server. The same recipes are MCP prompts (prompts/list) on `https://api.openkrill.app/gateway/mcp`.
Connect the server at `https://cars.openkrill.app/mcp` in your MCP client settings (remote streamable HTTP, no key), then call its tools as normal tool calls. They are not command line programs: Bash, curl and scripts cannot reach them.

## Decode a VIN

Job: Decode a VIN, with the Car Check MCP server.

Use the `decode_vin` tool for this job.

Work on the user's request. If there is none yet, ask what they need. A request this job handles: "What car is VIN 1HGCM82633A004353?"

Steps:
1. Call `decode_vin` as a normal MCP tool call, with the request's values in the fields of its inputSchema. It needs `vin`.
2. If a value the tool needs is missing and you cannot tell it from the request, ask the user one short question first.
3. This input is known to work; copy its format, not its values: `{"vin":"1HGCM82633A004353"}`
4. For the request "What car is VIN 1HGCM82633A004353?", a correct answer looks like this: status ok with the model year, make Honda, model, body class, engine, fuel and plant country, plus NHTSA's error_codes and error_text.
5. Follow the server's own rules:
   - Use these tools for questions about US cars by VIN or by year, make and model.
   - Call decode_vin to turn a VIN into the vehicle's details and surface any error_codes it returns, list_recalls for NHTSA recalls, get_fuel_economy for EPA mpg figures, and get_vehicle_report for recalls, crash ratings, complaints and fuel economy together.
   - Recalls and complaints are listed by year, make and model, not by VIN, so say so and point to nhtsa.gov/recalls to check one specific car.
   - Complaints are reports from owners, not confirmed defects.
   - A count of zero or a missing rating is not proof of safety.
   - Name every vehicle variant the data returns instead of picking one.
   - This is a data lookup, not a valuation, history report or inspection: do not give safety, mechanical, valuation or buying advice.
   - US vehicles only.
   - Credit NHTSA and the EPA.
6. If the call fails or the answer looks wrong, tell the user, then report it with `submit_feedback` (kind bug, tool `decode_vin`).

## Get a vehicle report

Job: Get a vehicle report, with the Car Check MCP server.

Use the `get_vehicle_report` tool for this job.

Work on the user's request. If there is none yet, ask what they need. A request this job handles: "Give me a report on a 2016 Honda Civic."

Steps:
1. Call `get_vehicle_report` as a normal MCP tool call, with the request's values in the fields of its inputSchema.
2. If a value the tool needs is missing and you cannot tell it from the request, ask the user one short question first.
3. This input is known to work; copy its format, not its values: `{"year":2016,"make":"Honda","model":"Civic"}`
4. For the request "Give me a report on a 2016 Honda Civic.", a correct answer looks like this: status ok with crash ratings per variant, a recall count with the newest recalls, a complaint count with the components named most often, and fuel economy, plus the notice that recalls and complaints are by year, make and model and that complaints are not confirmed defects.
5. Follow the server's own rules:
   - Use these tools for questions about US cars by VIN or by year, make and model.
   - Call decode_vin to turn a VIN into the vehicle's details and surface any error_codes it returns, list_recalls for NHTSA recalls, get_fuel_economy for EPA mpg figures, and get_vehicle_report for recalls, crash ratings, complaints and fuel economy together.
   - Recalls and complaints are listed by year, make and model, not by VIN, so say so and point to nhtsa.gov/recalls to check one specific car.
   - Complaints are reports from owners, not confirmed defects.
   - A count of zero or a missing rating is not proof of safety.
   - Name every vehicle variant the data returns instead of picking one.
   - This is a data lookup, not a valuation, history report or inspection: do not give safety, mechanical, valuation or buying advice.
   - US vehicles only.
   - Credit NHTSA and the EPA.
6. If the call fails or the answer looks wrong, tell the user, then report it with `submit_feedback` (kind bug, tool `get_vehicle_report`).

## List recalls

Job: List recalls, with the Car Check MCP server.

Use the `list_recalls` tool for this job.

Work on the user's request. If there is none yet, ask what they need. A request this job handles: "Any recalls on a 2018 Toyota Camry?"

Steps:
1. Call `list_recalls` as a normal MCP tool call, with the request's values in the fields of its inputSchema. It needs `year`, `make`, `model`.
2. If a value the tool needs is missing and you cannot tell it from the request, ask the user one short question first.
3. This input is known to work; copy its format, not its values: `{"year":2018,"make":"Toyota","model":"Camry"}`
4. For the request "Any recalls on a 2018 Toyota Camry?", a correct answer looks like this: status ok with the total number of recalls and, newest first, each recall's campaign number, date, component, summary, consequence and remedy, with the reminder that recalls are listed by model, not VIN.
5. Follow the server's own rules:
   - Use these tools for questions about US cars by VIN or by year, make and model.
   - Call decode_vin to turn a VIN into the vehicle's details and surface any error_codes it returns, list_recalls for NHTSA recalls, get_fuel_economy for EPA mpg figures, and get_vehicle_report for recalls, crash ratings, complaints and fuel economy together.
   - Recalls and complaints are listed by year, make and model, not by VIN, so say so and point to nhtsa.gov/recalls to check one specific car.
   - Complaints are reports from owners, not confirmed defects.
   - A count of zero or a missing rating is not proof of safety.
   - Name every vehicle variant the data returns instead of picking one.
   - This is a data lookup, not a valuation, history report or inspection: do not give safety, mechanical, valuation or buying advice.
   - US vehicles only.
   - Credit NHTSA and the EPA.
6. If the call fails or the answer looks wrong, tell the user, then report it with `submit_feedback` (kind bug, tool `list_recalls`).