dev-workflow

The dev-workflow skill provides a structured, agentic framework for executing the complete software development lifecycle (SDLC) within AI-assisted development environments. Designed for developers and engineers building production-ready applications, this skill transitions projects from initial ideation to deployment using a standardized, tool-driven methodology. It leverages Model Context Protocol (MCP) tools to interact with external services like Linear for issue tracking, Notion for documentation, GitHub for version control, and cloud providers for infrastructure, ensuring that every task is tracked, tested, and validated before reaching production.

• Orchestrates the full development lifecycle: Intake, Triage, Exploration, Planning, Coding, Testing, Code Review, Committing, Deploying, and Improvement.

• Integrates with MCP tools including mcp__linear, mcp__notion, mcp__github, and mcp__memory for context-aware task execution.

• Automates infrastructure management using Pulumi for Go-based IaC, ensuring reproducible deployments on Google Cloud Run and Firebase.

• Implements rigorous testing protocols including Playwright for E2E browser tests, Postman/Newman for API validation, and unit tests via standard test runners.

• Supports multi-environment workflows with automated CI/CD pipelines through GitHub Actions, handling security scans, linting, and health checks.

• Features decision-making logic for triage: OpenSpec for complex architectural proposals and TodoWrite for bug fixes and routine tasks.

• Users should initiate the workflow using natural language triggers like start, implement, or work on to begin a session.

• The agent relies on inputs such as issue IDs from Linear or context pages from Notion to ground its planning phase.

• Always capture learnings into the Memory store to avoid repetitive debugging and to improve future agentic performance.

• Ensure all code changes undergo standard review processes, including security scanning for auth or data-handling logic, before triggering deployment tasks.

• Constraints: The skill assumes a modern stack (Go, React/Next.js, Swift, Kotlin, Pulumi, Firebase) and requires pre-configured GCP/Firebase projects to execute deployments successfully.

• Follow the defined anti-patterns: avoid coding without prior exploration, skip Linear tracking, or submitting monolithic commits that complicate review cycles.