Roadmap
Plan and execute entire application builds. Generates phased delivery roadmaps, then executes them autonomously — phase by phase, committing at milest
Category: development Source: jezweb/claude-skillsWhat Is Roadmap?
The Roadmap skill for Claude Code is a powerful development tool designed to plan and autonomously execute entire application builds. Unlike broad strategy documents, Roadmap generates a detailed, stepwise delivery blueprint where each phase consists of actionable, concrete tasks. It enables both the planning and execution of complex software projects, supporting multi-phase delivery, autonomous advancement through milestones, and robust progress tracking. Roadmap is especially valuable for projects that require structured organization and iterative delivery, ensuring that the path from inception to launch is clear, executable, and efficient.
Why Use Roadmap?
Building modern applications demands more than just coding expertise—it requires thoughtful planning, phased execution, and continuous adaptation. The Roadmap skill addresses common pain points such as vague initial requirements, lack of structured progression, and the challenge of resuming work across sessions.
Key benefits include:
- End-to-End Planning: Translates research briefs or high-level ideas into a stepwise, actionable plan.
- Autonomous Execution: Enables Claude Code to proceed through the development lifecycle, automatically committing, deploying, and testing at each milestone.
- Reduced Backtracking: Phases are explicitly ordered, minimizing the risk of rework or missed dependencies.
- Persistent Progress: Supports long-term, multi-session projects, allowing work to pause and resume seamlessly.
This skill is ideal for developers, teams, and technical leads who want to streamline their application delivery process, ensure clarity in task delegation, and maintain momentum throughout complex builds.
How to Get Started
To utilize the Roadmap skill, you must provide a suitable project input, typically a deep research brief or a comprehensive product specification. The process can be broken down as follows:
- Prepare Your Input:
- Place your deep research brief or specification in a format accessible to Claude Code (e.g.,
.jez/artifacts/research.md).
- Place your deep research brief or specification in a format accessible to Claude Code (e.g.,
- Invoke the Skill:
- Use any of the supported triggers, such as
roadmap,plan the build, orbuild the whole thing, to initiate the planning phase.
- Use any of the supported triggers, such as
- Review the Generated Roadmap:
- Claude Code will output a multi-phase plan, where each phase lists concrete, actionable tasks.
- Start Execution:
- Use the
startmode (start building) to begin autonomous execution. Claude Code will proceed phase by phase, committing code, deploying, and testing as defined.
- Use the
- Monitor and Manage Progress:
- Check your current status with triggers like
statusorwhat phase are we on.
- Check your current status with triggers like
- Pause and Resume:
- If you need to interrupt development, use the
resumemode (resume the build) to continue later, picking up exactly where you left off.
- If you need to interrupt development, use the
Example: Triggering Roadmap Generation in a Claude Code Session
## Assuming your research brief is ready:
roadmap .jez/artifacts/research.md
Example: Starting Autonomous Execution
start building
Key Features
- Phased Delivery Planning: Each roadmap is divided into logical phases—design, setup, implementation, testing, deployment, and post-launch—each with actionable, granular tasks.
- Autonomous Phase Execution: Claude Code carries out tasks phase by phase, making commits at milestone boundaries, running tests, and handling deployments as specified.
- Flexible Operating Modes: Four primary modes—
plan,start,resume, andstatus—let you plan, begin, pause, and monitor builds seamlessly. - Trigger-Based Control: Natural language triggers make it easy to interact with the skill and manage build progress.
- Integration with Claude Code Tools: Leverages file operations (
Read,Write,Edit), shell execution (Bash), and code search utilities (Glob,Grep) for comprehensive automation.
Example: Example Roadmap Output
## Phase 1: Project Setup
- Initialize Git repository
- Set up base project structure
- Configure CI pipeline
## Phase 2: Core Feature Implementation
- Implement user authentication
- Build API endpoints for data management
- Write and run unit tests
## Phase 3: Deployment & QA
- Configure deployment scripts
- Deploy to staging environment
- Run integration tests
Best Practices
- Start with Comprehensive Inputs: The quality of the roadmap is directly tied to the quality of your research brief or specification. Invest time in capturing requirements, edge cases, and business logic upfront.
- Review Each Phase: While Roadmap aims for autonomy, it's beneficial to review each phase for accuracy, dependencies, and completeness—especially for unfamiliar or critical projects.
- Use Status Frequently: Regularly check progress using the
statusmode to ensure milestones are being met and adjust as needed. - Leverage Version Control: Ensure that all code changes are version-controlled. Roadmap commits at each milestone, but maintaining your own backups or branches is recommended for safety.
- Iterate When Necessary: If requirements change or issues arise, update your input brief and re-trigger the roadmap process for adaptive planning.
Important Notes
- Compatibility: The Roadmap skill is designed exclusively for Claude Code and is not compatible with other automation or AI platforms.
- Skill Limitations: While robust, Roadmap may encounter limitations in ambiguous or under-specified projects. Clear, actionable inputs are critical for optimal results.
- Autonomy Scope: Although Roadmap can execute autonomously, human oversight is advised—especially for deployments, production changes, or critical tasks.
- Error Handling: If Claude Code becomes "stuck," review logs and phase outputs to diagnose issues, then adjust your inputs or manually intervene as needed.
- Security Considerations: Always audit generated code, deployment scripts, and CI/CD configurations for security best practices before production use.
For more details and source code, visit the official repository.