> ## Documentation Index
> Fetch the complete documentation index at: https://docs.mixus.ai/llms.txt
> Use this file to discover all available pages before exploring further.

# Task Preparation

> Learn how to write effective evaluation tasks with ready-to-use templates

## Writing effective tasks

Good task descriptions lead to better results. Follow these guidelines:

### Be specific and clear

<CodeGroup>
  ```text Good Example theme={null}
  Calculate 15% commission on a $50,000 sale and send the result 
  via email to manager@company.com with subject "Q4 Commission Report"
  ```

  ```text Bad Example theme={null}
  Calculate commission and send email
  ```
</CodeGroup>

### Include expected outcomes

<CodeGroup>
  ```text Good Example theme={null}
  Research the current Bitcoin price from CoinMarketCap and calculate 
  the value of 0.5 BTC. Expected result should be around $20,000-$25,000.
  ```

  ```text Bad Example theme={null}
  Check crypto prices
  ```
</CodeGroup>

### Provide necessary context

<CodeGroup>
  ```text Good Example theme={null}
  Calculate R&D tax credit using Alternative Simplified Credit method 
  (IRS Form 6765 Section B). Current year QRE: $120,000. Prior 3-year 
  average: $90,000. Formula: 14% × (current_qre - base_amount)
  ```

  ```text Bad Example theme={null}
  Calculate tax credit
  ```
</CodeGroup>

***

## Email format templates

### Template 1: Auto-detect (Recommended)

```
To: agent@mixus.com
Subject: Eval: [Your Task Name]

[Clear description of what you want done]

Test type: with-verification
Auto-detect: true
Reviewer: your-email@company.com
```

**Example:**

```
To: agent@mixus.com
Subject: Eval: Market Research

Research the top 3 AI agent platforms by market share.
For each, find: pricing model, key features, and target market.

Test type: with-verification
Auto-detect: true
Reviewer: analyst@company.com
```

### Template 2: Manual checkpoints

```
To: agent@mixus.com
Subject: Eval: [Your Task Name]

[Task description]

Checkpoints:
1. [First verification point]
2. [Second verification point]
3. [Third verification point]

Test type: with-verification
Reviewer: your-email@company.com
```

**Example:**

```
To: agent@mixus.com
Subject: Eval: Financial Calculation

Calculate the R&D tax credit for $120,000 in qualified expenses.

Checkpoints:
1. Verify the base amount calculation is correct
2. Verify the final tax credit amount
3. Confirm the IRS form is properly filled

Test type: with-verification
Reviewer: cfo@company.com
```

### Template 3: Baseline (no verification)

```
To: agent@mixus.com
Subject: Eval: [Your Task Name]

[Task description]

Test type: without-verification
Reviewer: your-email@company.com
```

**Example:**

```
To: agent@mixus.com
Subject: Eval: Simple Calculation

Calculate 15% of $50,000

Test type: without-verification
Reviewer: test@company.com
```

***

## API format templates

### Template 1: Auto-detect (Recommended)

```json theme={null}
{
  "taskName": "Market Research",
  "taskDescription": "Research top 3 AI agent platforms, find pricing, features, and target market",
  "autoDetectCheckpoints": true,
  "testMode": "with-verification",
  "assignedReviewer": "analyst@company.com"
}
```

### Template 2: Manual checkpoints

```json theme={null}
{
  "taskName": "Financial Calculation",
  "taskDescription": "Calculate R&D tax credit for $120,000 QRE",
  "checkpoints": [
    {
      "stage": "base_calculation",
      "description": "Calculate base amount",
      "verificationQuestion": "Is the base amount $45,000?"
    },
    {
      "stage": "final_calculation",
      "description": "Calculate final tax credit",
      "verificationQuestion": "Is the tax credit $10,500?"
    }
  ],
  "testMode": "with-verification",
  "assignedReviewer": "cfo@company.com"
}
```

### Template 3: With webhook callback

```json theme={null}
{
  "taskName": "Automated Test",
  "taskDescription": "Calculate 15% commission on $50,000 sale",
  "autoDetectCheckpoints": true,
  "testMode": "with-verification",
  "assignedReviewer": "manager@company.com",
  "webhookUrl": "https://your-system.com/webhook/eval-complete",
  "externalId": "test-001"
}
```

***

## Task categories and examples

### Financial calculations

<Accordion title="Commission Calculation">
  ```
  Subject: Eval: Sales Commission Q4

  Calculate commission for Q4 sales:

  - Base salary: $60,000/year
  - Commission rate: 3% of sales over $100,000
  - Q4 sales: $450,000

  Calculate total Q4 compensation.

  Test type: with-verification
  Auto-detect: true
  Reviewer: finance@company.com

  ```
</Accordion>

<Accordion title="ROI Analysis">
  ```

  Subject: Eval: Marketing ROI

  Calculate ROI for Q3 marketing campaign:

  - Total spend: $75,000
  - Revenue generated: $225,000
  - Campaign duration: 3 months

  Provide ROI percentage and monthly breakdown.

  Test type: with-verification
  Auto-detect: true
  Reviewer: cmo@company.com

  ```
</Accordion>

### Research tasks

<Accordion title="Competitor Analysis">
  ```

  Subject: Eval: Competitor Research

  Research OpenAI ChatGPT, Anthropic Claude, and Google Gemini.
  For each, find:

  1. Pricing (all tiers)
  2. Key features
  3. API availability
  4. Target market

  Create comparison table.

  Test type: with-verification
  Auto-detect: true
  Reviewer: product@company.com

  ```
</Accordion>

<Accordion title="Market Trends">
  ```

  Subject: Eval: AI Market Trends

  Research current AI agent market trends:

  1. Market size and growth rate
  2. Top 5 companies by market share
  3. Emerging technologies
  4. Future outlook (2-year)

  Summarize findings in 2-page report.

  Test type: with-verification
  Auto-detect: true
  Reviewer: strategy@company.com

  ```
</Accordion>

### Communication tasks

<Accordion title="Customer Follow-up">
  ```

  Subject: Eval: Customer Email Draft

  Draft follow-up email to customer who requested product demo.
  Include:

  - Thank them for interest
  - Propose 3 time slots next week (9am, 2pm, 4pm EST)
  - Link to product features page
  - Professional, warm tone

  Test type: with-verification
  Auto-detect: true
  Reviewer: sales@company.com

  ```
</Accordion>

<Accordion title="Meeting Summary">
  ```

  Subject: Eval: Meeting Summary Email

  Create meeting summary email for quarterly review:

  - Key decisions made
  - Action items with owners
  - Next meeting date
  - Send to: team@company.com

  Test type: with-verification
  Auto-detect: true
  Reviewer: manager@company.com

  ```
</Accordion>

***

## Best practices

### Do's ✅

<Check>**Be specific about numbers and amounts**</Check>

```

Good: "Calculate 15% commission on $50,000"
Bad: "Calculate some commission"

```

<Check>**Include verification criteria**</Check>

```

Good: "Expected result should be $7,500 (15% of $50,000)"
Bad: "Calculate it"

```

<Check>**Specify data sources when relevant**</Check>

```

Good: "Get Bitcoin price from CoinMarketCap"
Bad: "Get crypto price"

```

<Check>**Break complex tasks into steps**</Check>

```

Good: "1) Research competitors 2) Analyze pricing 3) Create table 4) Email summary"
Bad: "Do competitor analysis"

```

### Don'ts ❌

<Warning>**Don't use vague language**</Warning>

```

Bad: "Do some research about AI"
Good: "Research top 3 AI companies: OpenAI, Anthropic, Google DeepMind"

```

<Warning>**Don't forget to specify reviewer**</Warning>

```

Bad: Missing "Reviewer:" field
Good: "Reviewer: manager@company.com"

```

<Warning>**Don't mix manual checkpoints with auto-detect**</Warning>

```

Bad: Both "Checkpoints: 1, 2, 3" AND "Auto-detect: true"
Good: Choose ONE approach

```

***

## Checkpoint placement guide

### AI auto-detects checkpoints before:

<CardGroup cols={2}>
  <Card title="External Communications" icon="envelope">
    * Sending emails
    * Posting to social media
    * Sending SMS/messages
  </Card>

  <Card title="Financial Actions" icon="dollar-sign">
    * Making purchases
    * Transferring money
    * Processing payments
  </Card>

  <Card title="Irreversible Operations" icon="trash">
    * Deleting data
    * Canceling subscriptions
    * Submitting forms
  </Card>

  <Card title="High-Value Decisions" icon="gavel">
    * Strategic recommendations
    * Large calculations
    * Important approvals
  </Card>
</CardGroup>

### AI does NOT checkpoint for:

* Web searches and research
* Reading files or data
* Analysis and calculations (unless result triggers action)
* Information gathering
* Report generation (unless sending/publishing)

***

## Field reference

### Required fields

| Field                         | Description       | Example                                             |
| ----------------------------- | ----------------- | --------------------------------------------------- |
| `taskName` / Subject          | Task identifier   | "Calculate Commission"                              |
| `taskDescription` / Body      | What to do        | "Calculate 15% of \$50,000"                         |
| `testMode` / Test type        | Verification mode | "with-verification"                                 |
| `assignedReviewer` / Reviewer | Who verifies      | "[manager@company.com](mailto:manager@company.com)" |

### Optional fields

| Field                                 | Description                | Example                                                        |
| ------------------------------------- | -------------------------- | -------------------------------------------------------------- |
| `autoDetectCheckpoints` / Auto-detect | Let AI find checkpoints    | true                                                           |
| `checkpoints` / Checkpoints           | Manual verification points | See checkpoint format below                                    |
| `webhookUrl`                          | Callback URL               | "[https://your-app.com/webhook](https://your-app.com/webhook)" |
| `externalId`                          | Your tracking ID           | "task-001"                                                     |

***

## Common mistakes

<AccordionGroup>
  <Accordion title="Mistake: Subject doesn't start with 'Eval:'">
    **Wrong:** `Subject: Calculate Commission`
    **Right:** `Subject: Eval: Calculate Commission`
  </Accordion>

  <Accordion title="Mistake: Vague task description">
    **Wrong:** `Do some calculations`
    **Right:** `Calculate 15% commission on a $50,000 sale`
  </Accordion>

  <Accordion title="Mistake: Missing reviewer">
    **Wrong:** No reviewer specified
    **Right:** `Reviewer: manager@company.com`
  </Accordion>

  <Accordion title="Mistake: Too many checkpoints">
    **Wrong:** 10 checkpoints for simple task
    **Right:** 2-4 checkpoints for most tasks, or use auto-detect
  </Accordion>
</AccordionGroup>

***

## Next steps

<CardGroup cols={2}>
  <Card title="See Examples" icon="lightbulb" href="/evaluation/examples">
    10+ ready-to-use example tasks
  </Card>

  <Card title="API Reference" icon="code" href="/eval-api">
    Complete API documentation
  </Card>

  <Card title="Best Practices" icon="star" href="/evaluation/best-practices">
    Advanced tips and tricks
  </Card>

  <Card title="Get Started" icon="rocket" href="/evaluation/getting-started">
    Submit your first task
  </Card>
</CardGroup>

```
```