---
allowed-tools: [Bash, Read, Write, Glob]
description: Starts a new or updated MRD (Market Requirements Document) research session.
---

# Start MRD Research Session

## Context
- **User Request:** $ARGUMENTS
- **MRD Directory:** `.taskmaster/docs/mrd/`
- **Existing Sessions:** !`ls -ld .taskmaster/docs/mrd/*/ 2>/dev/null || echo "No existing sessions found"`

## Goal
To initialize a new or updated Market Requirements Document (MRD) research session, set up the dedicated workspace, and present the first research task to the user.

## Process
1.  **Determine Session Index:**
    - Scan the `.taskmaster/docs/mrd/` directory to find the highest existing session index.
    - Assign the next sequential number for the new session.

2.  **Create Session Directory:**
    - Create a new directory named `[index]-[session_name]` inside `.taskmaster/docs/mrd/`.
    - Example: `.taskmaster/docs/mrd/001-mvp-launch/`

3.  **Handle Update Session (if `--from` is provided):**
    - Copy all `0*_*.md` user research files from the base session directory to the new session directory.
    - Generate a `_00_update_kickoff_report.md` file. This report will compare the goals of the base session and the new session, highlighting key assumptions that need to be re-validated.

4.  **Initialize Session State:**
    - Create a `_session-state.json` file in the new session directory.
    - Initialize it with session details (index, name, status: 'initialized', etc.).

5.  **Interactive Hypothesis Definition:**
    -   This step is a structured conversation to build the foundational `01_initial_hypothesis.md`.

    -   **For a new session:**
        -   The AI will ask a series of clarifying questions to build the hypothesis, such as:
            1.  "What is the core problem you are trying to solve?"
            2.  "Who is the primary target audience for this product?"
            3.  "At a high level, what is your proposed solution?"
            4.  "What is the unique value proposition? Why will users choose it over alternatives?"
        -   After gathering the user's answers, the AI will synthesize them into a coherent initial hypothesis.
        -   The AI saves this synthesized content into the `01_initial_hypothesis.md` file.

    -   **For an update session (`--from` is used):**
        -   The AI first presents the key findings from the generated `_00_update_kickoff_report.md`.
        -   It then asks for the user's input on the re-validation points. For example: "The report suggests we need to re-validate our target audience. Has your understanding of the target customer changed? If so, how?"   
        -   The user's feedback is incorporated to refine the session's starting assumptions, which can be noted in the kickoff report or a new hypothesis file.

6.  **Assign First Task and Generate Research Prompt:**
    -   Based on the session's goal and initial hypothesis, the AI selects the most logical first research task from the examples below.
    -   It clearly presents this task to the user, specifying the filename for the output.
    -   **Crucially**, it must also generate a detailed, self-contained research prompt that can be used by any external agent or tool. This prompt should be presented to the user in a structured format.

    **Research Prompt Generation Template:**
    The AI will use the newly created `01_initial_hypothesis.md` to construct a prompt like this:
    ```
    ### Research Prompt: [Objective of the Task]

    **1. Project Context:**
       - **Product/Idea:** [Synthesized from 01_initial_hypothesis.md - e.g., "A platform connecting eco-friendly suppliers with small businesses."]
       - **Core Problem:** [From hypothesis - e.g., "Small businesses struggle to find and verify sustainable suppliers."]
       - **Target Audience:** [From hypothesis - e.g., "Owners of small to medium-sized retail businesses."]
       - **Proposed Solution:** [From hypothesis - e.g., "A curated, searchable marketplace with a built-in verification system."]

    **2. Research Objective:**
       - [A clear goal for this specific research task, e.g., "To deeply understand the competitive landscape for our proposed solution."]

    **3. Key Research Questions:**
       - [A list of 3-5 specific questions. For competitor analysis, it could be:]
         - Who are the top 3 direct and indirect competitors?
         - What are their pricing models and key features?
         - What are their primary strengths and weaknesses (SWOT)?
         - What market segment do they primarily target?

    **4. Expected Deliverables:**
         - A summary of findings.
         - Detailed answers to each key research question.
         - A concluding analysis of opportunities and threats for our product.
    ```

    **First Research Task Examples:**
    -   **Task:** "Define the target market size, segments, and create detailed user personas."
        -   **Filename:** `02_market_and_persona.md`
    -   **Task:** "Validate the core problem this product aims to solve and outline the proposed solution's unique value proposition."
        -   **Filename:** `02_problem_and_solution.md`
    -   **Task:** "Identify the top 3-5 direct and indirect competitors and analyze their strengths, weaknesses, and market positioning."
        -   **Filename:** `02_competitive_landscape.md`
    -   **Task:** "Brainstorm a list of potential core features that address the initial hypothesis and align with the target user's needs."
        -   **Filename:** `02_initial_feature_ideas.md`
    -   **Task:** "Define the key success metrics and Key Performance Indicators (KPIs) that will be used to measure the product's success."
        -   **Filename:** `02_success_metrics.md`

    -   After presenting the task and the detailed prompt, the AI will conclude with: "Once you have completed your research and saved it to the specified file, please run the `/planning/mrd/2-analyze-research-data` command to proceed."

## Example Usage
-   **Start a new session:**
    `/planning/mrd/1-start-session --name="mvp-launch"`
-   **Start a session based on a previous one:**
    `/planning/mrd/1-start-session --name="enterprise-expansion" --from="mvp-launch"`