regardv/legendaryteam_for_claude
25 subagents
Subagent
architecture-agentArchitecture design and planning specialist requiring human approvalbug-resolverBug investigation, root cause analysis, and systematic resolutionchiefMaster orchestrator for parallel autonomous teams with confidence-based routingcodebase-cartographerArchitecture guardian maintaining codebase map and tracking all file changesconfidence-agentAnalyzes tasks, calculates confidence scores, and routes to appropriate execution tierdatabase-agentDatabase schema, migration, and CRUD operations specialistdiscovery-protectorDrift detection guardian that blocks execution if code diverges from specsdoc-agentDocumentation creation and maintenance specialiste2e-runnerPlaywright-based E2E test generation, execution, and maintenanceinfra-guardianInfrastructure validator that blocks deployments on config driftinfrastructure-agentInfrastructure, deployment, and scaling specialist requiring human approvalopenspec-policeSpec enforcement officer banning chat TODOs and enforcing OpenSpec disciplineperformance-optimizerPerformance engineering specialist for profiling, optimization, and benchmarkingplannerDecomposes high-level goals into structured execution plans with dependenciesproject-analyzerDeep scanner analyzing project health and detecting technical debtrefactor-agentCode refactoring and optimization specialistreflection-agentEvaluates quality of actions and provides feedback for continuous improvementsecurity-agentSecurity implementation and audit specialist requiring human approvalsession-orchestratorSession memory and continuity guardian managing state across context clearsspecarchitectOpenSpec master managing backups, recompilation, and spec-code synchronizationteam-builderDynamic agent factory that rebuilds team based on detected technology stacktechstack-fingerprinterTechnology stack detector that scans codebase and outputs tech_stack.yamltest-agentAutonomous test generation and quality assurance specialistui-agentUI component creation and styling specialistverifierReviews plans and output for completeness, errors, and edge cases