sygaldry: Command for Claude Code

Analyze git working state and create comprehensive Linear issues using the Linear MCP

1. Check Linear Context:

   • Use mcp__linear__get_team to get team info
   • Use mcp__linear__list_issue_statuses to get available statuses
   • Use mcp__linear__list_issue_labels to get available labels
   • Use mcp__linear__list_projects to identify relevant project

2. Search for Related Existing Issues and Projects:

   • Use mcp__linear__list_projects to find active projects that could organize this work
   • Check for projects with status != "completed" or "cancelled"
   • Use mcp__linear__list_issues to find related open issues
   • Check issue titles and descriptions for relevance
   • Determine if changes should be attached to existing issues or projects

3. Categorize Changes:

   • Group by component/module
   • Identify change types and map to Linear labels:
     • Feature (new functionality)
     • Bug (fixes)
     • Refactor (code improvements)
     • Documentation (docs updates)
     • Improvement (enhancements)
     • Testing (test additions/updates)
     • Technical Debt (cleanup)
     • Performance (optimizations)
     • Security (security improvements)
     • Maintenance (dependency updates, etc.)

4. Create Project and Issue Structure:

   For major groups of related changes:

   a) Find or Create a Project to organize related work:

   // First, check for existing applicable projects
   const existingProjects = await mcp__linear__list_projects({ 
     teamId: team.id,
     includeArchived: false 
   });
   
   // Look for active projects that match our work
   let project = existingProjects.find(p => {
     const isActive = !p.completedAt && !p.canceledAt && !p.archivedAt;
     const isRelevant = p.name.includes("Refactor") || 
                       p.name.includes("AI Agent") ||
                       p.description?.includes("agent") ||
                       p.description?.includes("refactor");
     return isActive && isRelevant;
   });
   
   // Only create new project if no suitable one exists
   if (!project) {
     project = await mcp__linear__create_project({
       name: "Refactor: AI Agent Architecture",
       teamId: team.id,
       summary: "Consolidate AI agent system and remove redundant code",
       description: `## Overview
       Major refactoring effort to improve the AI agent system by removing redundant code and implementing cleaner patterns.
     
     ## Scope
     - Remove duplicate response models across all agents
     - Consolidate tool imports
     - Fix import issues and circular dependencies
     - Implement proper factory pattern
     - Clean up unused evaluation files
     
     ## Affected Components
     - Healthcare Agent
     - Education Agent  
     - Exa Agent
     - Agent Registry and Factory
     - Test Suite
     
     ## Technical Approach
     1. Audit current agent implementations
     2. Remove duplicate files
     3. Update imports to use centralized tools
     4. Ensure all tests pass
       5. Update documentation`,
       targetDate: "2025-07-01"
     });
   } else {
     console.log(`Using existing project: ${project.name} (${project.id})`);
   }