Help us improve
Share bugs, ideas, or general feedback.
Share bugs, ideas, or general feedback.
Share bugs, ideas, or general feedback.
Multiagent orchestration plugin for enterprise code development with specialized agents and skills, beads integration, and quality gates enforcement
npx claudepluginhub joshuarweaver/cascade-ai-ml-agents-misc-1 --plugin ivankristianto-k2-devCheck system prerequisites and project configuration
Start planning workflow to analyze requirements and create actionable beads tasks
Analyze and update a beads ticket with technical analysis and refined requirements
Start implementation workflow for one or more beads tickets
Create comprehensive test plan and test cases for a beads ticket
Use this agent when implementing features, writing code based on plans from beads tasks, performing self-review against quality gates, responding to code review feedback, or executing implementation work. This is the implementation specialist who writes code following approved plans and quality standards. The Engineer completes implementation and pushes changes, then the Technical Lead creates the PR. Examples: <example>Context: /k2:start command has completed setup and is launching the Engineer for implementation. user: "The /k2:start command has set up the worktree for beads-123. Please implement the authentication feature according to the plan." assistant: "I'll use the engineer agent to implement the authentication feature following the plan in beads-123." <commentary>The Engineer is launched by the /k2:start command (which runs Technical Lead logic directly) with a prepared worktree and task context, making this the primary triggering scenario for the engineer agent.</commentary></example> <example>Context: User wants to start implementation work on a beads task that has a plan. user: "Work on beads-456 to implement the user profile feature." assistant: "I'll use the engineer agent to implement the user profile feature according to the plan in beads-456." <commentary>When a user directly requests implementation work on a beads task, the engineer agent should be invoked to handle the implementation and self-review workflow.</commentary></example> <example>Context: Reviewer has provided feedback on a PR and changes are needed. user: "The reviewer left feedback on PR #789. Can you address the comments?" assistant: "I'll use the engineer agent to address the review feedback on PR #789." <commentary>The Engineer handles review feedback iterations (up to 2 cycles) before creating follow-up tickets, making this a clear engineering responsibility.</commentary></example> <example>Context: Planning is complete and implementation is ready to begin. user: "Continue with beads-234 implementation now that the plan is approved." assistant: "I'll use the engineer agent to implement beads-234 following the approved plan." <commentary>After the planning phase is complete, the Engineer takes over to execute the implementation work.</commentary></example>
Use this agent when analyzing requirements to create implementation plans and beads tasks, designing features or changes, planning complex implementations, breaking down user stories into actionable tasks, or when the /k2:planner command is executed. The Planner transforms high-level requirements into structured, hierarchical beads tasks with proper dependencies. Examples: <example>Context: User has executed the /k2:planner command with a feature description. user: "/k2:planner Add user authentication with JWT and refresh tokens" assistant: "I'll use the planner agent to analyze the authentication requirements and create a comprehensive implementation plan with beads tasks." <commentary>The /k2:planner command explicitly triggers the Planner agent to begin the requirements analysis and planning workflow, including codebase exploration, clarification, and task creation.</commentary></example> <example>Context: User wants to plan a new feature before implementation. user: "I want to add a real-time notification system. Can you help me plan this out?" assistant: "I'll use the planner agent to analyze the notification system requirements, explore the codebase, and create a detailed implementation plan." <commentary>When a user requests planning for a feature using keywords like "plan", "design", or "add a feature", the Planner should be invoked to handle requirements analysis and structured planning.</commentary></example> <example>Context: Technical Lead requests planning for a complex feature. user: "We need to refactor the authentication system to support multiple providers. Let's plan this carefully." assistant: "I'll use the planner agent to analyze the current authentication architecture and create a comprehensive refactoring plan with tasks." <commentary>Complex features or architectural changes benefit from the Planner's systematic approach to requirements analysis, codebase exploration, and structured task breakdown.</commentary></example> <example>Context: User provides requirements that need clarification and breakdown. user: "Build a dashboard that shows analytics and has export functionality" assistant: "I'll use the planner agent to clarify the requirements and create a detailed implementation plan with task dependencies." <commentary>When requirements are somewhat vague or need exploration, the Planner uses its full codebase access to understand context and asks clarifying questions to create accurate plans.</commentary></example>
Create pull requests for completed work
Use this agent when performing code reviews on GitHub pull requests, validating code quality against AGENTS.md standards, checking for security vulnerabilities, providing actionable feedback, or approving PRs that meet quality gates. This is a read-only review specialist who validates code but does not make changes. Examples: <example>Context: Engineer has created a PR and needs code review. user: "The Engineer created PR #456 for beads-123. Can you review it?" assistant: "I'll use the reviewer agent to perform a comprehensive code review of PR #456." <commentary>The Reviewer is explicitly assigned a code review task for a PR created by the Engineer, making this the primary triggering scenario for the reviewer agent.</commentary></example> <example>Context: Technical Lead has assigned review work after implementation completes. user: "Review the pull request for beads-789 to ensure it meets our quality standards." assistant: "I'll use the reviewer agent to validate the PR against our quality gates and standards." <commentary>When a Technical Lead or user requests code review with quality validation, the Reviewer agent should be invoked to perform thorough review against project standards.</commentary></example> <example>Context: PR is ready for validation and feedback. user: "PR #234 is ready for review. Please check security and best practices." assistant: "I'll use the reviewer agent to review PR #234 with focus on security and best practices." <commentary>The Reviewer handles security validation, best practices checking, and comprehensive code review, making this a clear reviewer responsibility.</commentary></example> <example>Context: Engineer has addressed feedback and needs re-review (iteration 1 or 2). user: "I've addressed the review feedback on PR #567. Can you take another look?" assistant: "I'll use the reviewer agent to re-review PR #567 and verify the feedback has been addressed." <commentary>The Reviewer performs iterative reviews (up to 2 iterations) to validate that feedback has been properly addressed, continuing until approval or follow-up tickets are needed.</commentary></example>
Use this agent for technical architecture analysis, system design, architectural review, code analysis, technical debt assessment, and ticket refinement with engineering depth. Examples: <example>Context: User needs architecture review for a new feature. user: "Review the proposed auth system design" assistant: "I'll use the technical-lead agent to analyze the architecture and provide design feedback."</example> <example>Context: User wants to understand codebase architecture. user: "How is the data layer structured?" assistant: "I'll use the technical-lead agent to explore and explain the system architecture."</example> <example>Context: User needs ticket refined with technical details. user: "Refine beads-123 with implementation approach" assistant: "I'll use the technical-lead agent to analyze and update the ticket with technical analysis."</example> <example>Context: User wants code reviewed for architectural soundness. user: "Review this PR for design patterns" assistant: "I'll use the technical-lead agent to assess architectural quality and design adherence."</example>
This skill should be used when the user asks to "work with beads tasks", "create a beads ticket", "update beads status", "read task comments", "manage dependencies", "sync beads", "use bd commands", or needs guidance on beads task management workflows. Provides comprehensive beads CLI usage and workflow patterns.
This skill should be used when the user asks to "review code", "perform code review", "check code quality", "review PR", "provide code feedback", or needs guidance on code review best practices and standards in k2-dev workflows.
This skill should be used when the user asks to "create a worktree", "manage git worktrees", "work on multiple branches simultaneously", "isolate task work", "use bd worktree", or needs guidance on git worktree patterns for task isolation in k2-dev workflows.
This skill should be used when the user needs to plan features, break down requirements, create implementation tasks, or design software architecture. Use this skill for comprehensive requirements analysis, technical planning, and creating structured task hierarchies with dependencies. The skill executes in the main conversation context.
Quick PR creation reference
Uses power tools
Uses Bash, Write, or Edit tools
Share bugs, ideas, or general feedback.
Own this plugin?
Verify ownership to unlock analytics, metadata editing, and a verified badge.
Sign in to claimOwn this plugin?
Verify ownership to unlock analytics, metadata editing, and a verified badge.
Sign in to claimBased on adoption, maintenance, documentation, and repository signals. Not a security audit or endorsement.
AI-supervised issue tracker for coding workflows. Manage tasks, discover work, and maintain context with simple CLI commands.
Autonomous multi-agent development pipeline โ from ticket to PR in one command
Plan-only mode and GitHub-issue-to-Beads workflow. Provides planning without auto-execution and structured work decomposition into Beads epics, tasks, and sub-tasks with verifiable acceptance criteria.
Virtual development team: TDD, debugging, code review, backlog management, and proven workflow patterns
Comprehensive skill pack with 66 specialized skills for full-stack developers: 12 language experts (Python, TypeScript, Go, Rust, C++, Swift, Kotlin, C#, PHP, Java, SQL, JavaScript), 10 backend frameworks, 6 frontend/mobile, plus infrastructure, DevOps, security, and testing. Features progressive disclosure architecture for 50% faster loading.
Harness-native ECC plugin for engineering teams - 64 agents, 262 skills, 84 legacy command shims, reusable hooks, rules, MCP conventions, and operator workflows for Claude Code plus adjacent agent harnesses
Enterprise-grade multiagent orchestration plugin for Claude Code that manages end-to-end software development workflows with specialized AI agents.
k2-dev simulates a complete development team with specialized agents and skills:
Agents:
Skills (main conversation context):
bv command)gh command)AGENTS.md - Agent behavior guidelinesconstitution.md - Project principles and constraintsInstall via Claude Marketplace:
claude plugin install k2-dev@ivankristianto/k2-dev
/k2:startStart implementation workflow for one or more tickets.
Usage:
/k2:start beads-123 [--skip-worktree]
/k2:start beads-123,beads-234,beads-345 [--skip-worktree]
/k2:start [--skip-worktree] # No ticket - launches beads_viewer (bv) for interactive selection
Options:
--skip-worktree: Create branch in main repository instead of creating an isolated worktreeInteractive Mode:
When called without a ticket ID, /k2:start will launch the bv (beads_viewer) command for interactive task selection. This provides a visual interface to browse and select tasks from your beads project.
Workflow:
feature/beads-{id}) or creates branch in main repo if --skip-worktree is used/k2:plannerStart planning workflow for new features using the Planning skill.
Usage:
/k2:planner
Workflow:
Benefits:
/k2:reportGenerate status report for a ticket.
Usage:
/k2:report beads-123
Output: Structured markdown report with:
/k2:testCreate test plan for a ticket using the Test Planning skill.
Usage:
/k2:test beads-123
Workflow:
/k2:doctorCheck system prerequisites and project configuration.
Usage:
/k2:doctor
What it checks:
bd (beads), bv (beads_viewer), gh (GitHub CLI), and git with worktree support are installedAGENTS.md and constitution.md files.beads/ directory and beads operational statusbd compact if neededOutput: Diagnostic report showing which requirements are met, missing, or need attention, plus actionable recommendations for beads maintenance.