Files
claude-task-master/tasks/task_060.txt
2025-04-11 15:08:58 -04:00

74 lines
3.8 KiB
Plaintext

# Task ID: 60
# Title: Implement Mentor System with Round-Table Discussion Feature
# Status: pending
# Dependencies: None
# Priority: medium
# Description: Create a mentor system that allows users to add simulated mentors to their projects and facilitate round-table discussions between these mentors to gain diverse perspectives and insights on tasks.
# Details:
Implement a comprehensive mentor system with the following features:
1. **Mentor Management**:
- Create a `mentors.json` file to store mentor data including name, personality, expertise, and other relevant attributes
- Implement `add-mentor` command that accepts a name and prompt describing the mentor's characteristics
- Implement `remove-mentor` command to delete mentors from the system
- Implement `list-mentors` command to display all configured mentors and their details
- Set a recommended maximum of 5 mentors with appropriate warnings
2. **Round-Table Discussion**:
- Create a `round-table` command with the following parameters:
- `--prompt`: Optional text prompt to guide the discussion
- `--id`: Optional task/subtask ID(s) to provide context (support comma-separated values)
- `--turns`: Number of discussion rounds (each mentor speaks once per turn)
- `--output`: Optional flag to export results to a file
- Implement an interactive CLI experience using inquirer for the round-table
- Generate a simulated discussion where each mentor speaks in turn based on their personality
- After all turns complete, generate insights, recommendations, and a summary
- Display results in the CLI
- When `--output` is specified, create a `round-table.txt` file containing:
- Initial prompt
- Target task ID(s)
- Full round-table discussion transcript
- Recommendations and insights section
3. **Integration with Task System**:
- Enhance `update`, `update-task`, and `update-subtask` commands to accept a round-table.txt file
- Use the round-table output as input for updating tasks or subtasks
- Allow appending round-table insights to subtasks
4. **LLM Integration**:
- Configure the system to effectively simulate different personalities using LLM
- Ensure mentors maintain consistent personalities across different round-tables
- Implement proper context handling to ensure relevant task information is included
Ensure all commands have proper help text and error handling for cases like no mentors configured, invalid task IDs, etc.
# Test Strategy:
1. **Unit Tests**:
- Test mentor data structure creation and validation
- Test mentor addition with various input formats
- Test mentor removal functionality
- Test listing of mentors with different configurations
- Test round-table parameter parsing and validation
2. **Integration Tests**:
- Test the complete flow of adding mentors and running a round-table
- Test round-table with different numbers of turns
- Test round-table with task context vs. custom prompt
- Test output file generation and format
- Test using round-table output to update tasks and subtasks
3. **Edge Cases**:
- Test behavior when no mentors are configured but round-table is called
- Test with invalid task IDs in the --id parameter
- Test with extremely long discussions (many turns)
- Test with mentors that have similar personalities
- Test removing a mentor that doesn't exist
- Test adding more than the recommended 5 mentors
4. **Manual Testing Scenarios**:
- Create mentors with distinct personalities (e.g., Vitalik Buterin, Steve Jobs, etc.)
- Run a round-table on a complex task and verify the insights are helpful
- Verify the personality simulation is consistent and believable
- Test the round-table output file readability and usefulness
- Verify that using round-table output to update tasks produces meaningful improvements