63 specialized agents working together to build software.
Traditional AI assistants wait for you to type, respond once, and stop. An agentic AI is different—it can take actions on your behalf. It reads files, runs commands, browses the web, and iterates on its own work until the job is done.
This toolkit gives you a collection of 63 specialized agents, each designed for a specific development task. Some write code. Some review it. Some run tests or deploy infrastructure. When you work with a primary agent, it automatically calls on specialists as needed—so you get expert help without managing each step yourself.
Think of it as a team of focused collaborators, ready to help you build software faster and more reliably.
These are the entry points to the system. Start a conversation with any of them, and they'll orchestrate specialists as needed.
Builds features from PRDs or ad-hoc requests by orchestrating implementation agents
Refine draft PRDs and prepare them for implementation
Maintains the AI toolkit - agents, skills, templates, and scaffolds
You guide the direction. Agents handle the execution.
Work with @planner to refine requirements. You decide what gets built and how.
@builder implements your PRD. You verify, steer corrections, and approve commits.
Improve agents over time via @toolkit. You approve changes that ripple across projects.
The Process
From idea to shipped feature — and back again. Every feature follows this cycle.
Bugs feed back into the loop
Discovered issues can be tracked as new PRDs
Swipe to explore
63 Agents
Sub-agents that get delegated to automatically when primary agents need specialized help.
28
Review code for correctness, security, style, and best practices.
15
Implement features across different languages and frameworks.
9
Write and run unit tests, integration tests, and E2E tests.
3
Coordinate other agents and manage workflows.
8
Debugging, docs, cleanup, and specialized tools.
32 Skills
Skills are specialized instruction sets that agents load on-demand for complex, multi-step workflows. Unlike agents, skills provide situational expertise—loaded when needed, not always active.
adhoc-workflow
Ad-hoc mode workflow for Builder
agent-audit
Audit AI toolkit agents for compliance and coverage gaps
agent-onboard
Onboard new or modified agents to ensure they are project-context aware and follow established conventions
builder-state
Manage builder session state for resumability
cve
Assess a CVE for exposure, exploitability, and remediation options across one or more repositories
e2e-quality
Quality-beyond-correctness E2E testing patterns
marketing-copy
Generate marketing copy from product documentation and PRDs
merge-conflicts
Resolve merge conflicts on a pull request by merging the target branch into the source branch
Get up and running in a few minutes.
Get the OpenCode desktop app for your platform. It's the AI-native code editor that powers the agent system.
Download OpenCodeOpen the OpenCode command palette with Ctrl+P, type provider, and select "Connect Provider". Then choose how you want to connect to a model.
Open a terminal and run the toolkit installer:
curl -fsSL https://raw.githubusercontent.com/mdmagnuson-creator/yo-go/main/install.sh | bash