Pair Programming

Runs AI-assisted pair programming sessions with role switching, testing, and code review.

1|1|Updated Nov 28, 2025
One-click install
npx skills add https://github.com/33may/robotics --skill pair-programming-33may
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: Pair Programming
Source: https://github.com/33may/robotics/tree/main/humanoid/.claude/skills/pair-programming
Command: npx skills add https://github.com/33may/robotics --skill pair-programming-33may

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill requires claude-flow.

What problem does it solve? Solo development often lacks real-time feedback, structured testing discipline, and continuous code review, leading to bugs, security gaps, and inconsistent quality that surface late in the workflow. ## Core Features & Use Cases - Multiple Collaboration Modes: Driver, Navigator, Switch, TDD, Review, Mentor, and Debug modes with automatic role switching at configurable intervals. - Real-Time Verification: Truth-score quality monitoring with automatic rollback, security scanning, performance analysis, and coverage enforcement. - Testing Integration: Automatic test generation, coverage tracking, and continuous test execution during sessions. - Use Case: Start a TDD session to build a shopping cart feature where the AI writes failing tests, you implement minimal code to pass them, and the pair refactors together while coverage and truth scores are tracked. ## Quick Start Start a pair programming session in TDD mode with test-first enabled and a 90 percent coverage target.

Frequently Asked Questions about Pair Programming

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I start a pair programming session with claude-flow?

Run claude-flow pair --start to begin a basic session. Add flags like --mode tdd, --agent senior-dev, or --verify to customize the collaboration mode, AI partner, and quality verification level.

What is the difference between driver and navigator mode?

In driver mode you write code while the AI provides guidance and review. In navigator mode the AI writes implementation code while you provide direction and architectural decisions. Switch mode alternates roles at set intervals.

Does claude-flow pair support test-driven development?

Yes, TDD mode follows the red-green-refactor cycle with --test-first and --coverage flags. The AI generates failing tests, you implement minimal passing code, and the pair refactors while coverage thresholds are enforced.

What is the truth score in pair programming sessions?

The truth score is a quality metric from 0 to 1 that verifies code changes before commits. Scores below 0.90 trigger errors, 0.95-0.98 is good, and above 0.98 is excellent, with automatic rollback on failures when enabled.

Can I save and resume a pair programming session?

Yes, sessions support auto-save at configurable intervals, manual saving with --save, and recovery with --load or --recover. You can also export sessions and generate reports for completed work.

Why won't my pair programming session start?

Check that the claude-flow CLI is installed, verify your configuration file has valid JSON syntax, ensure the workspace is clean, and review log files. Run claude-flow pair config validate to diagnose configuration issues.