Oh My Opencode Features

Oh My OpenCode Features

Overview

This skill provides a comprehensive reference of all features available in oh-my-opencode, including curated agents, powerful tools, MCP integrations, hook systems, and productivity enhancements.

When to Use

Trigger phrases:

• "oh my opencode features"

• "Agent orchestration – When you need specialized agents for complex tasks, de"

• "MultiModel workflows – When leveraging multiple LLM providers, fallback chai"

• "Background execution – When running parallel agents, async task completion,"

• Agent orchestration – When you need specialized agents for complex tasks, deep work, architecture decisions, documentation, or codebase exploration

• MultiModel workflows – When leveraging multiple LLM providers, fallback chains, or per-agent model optimization

• Background execution – When running parallel agents, async task completion, or autonomous deep work

• LSP/AST tools – When refactoring, renaming, searching code with AST awareness, or running diagnostics

• Hooks integration – When automating workflows via PreToolUse, PostToolUse, UserPromptSubmit, Stop triggers

• Session management – When managing session history, session continuity, or session-based task continuation

• Workflow automation – When using Ultrawork, Ralph Loop, TDD, debugging patterns

• Quality verification – When enforcing comment checks, TODO enforcement, and verification before completion

Core Features

• Core operation execution with comprehensive error handling
• Input validation and output quality assurance
• Integration with existing workflows and toolchains
• Detailed logging for debugging and audit trails

1. Curated Agent System

Sisyphus — Main Orchestrator

• Purpose: Primary autonomous agent for complex tasks
• Key Features:
  • Goal-oriented execution without step-by-step instructions
  • Multi-model orchestration
  • Parallel agent launches (2-5 agents simultaneously)
  • Strict TODO list enforcement
  • Hash-anchored edit tool (LINE#ID format)
  • Comment checker for clean code
  • Claude Code compatibility layer

Hephaestus — Autonomous Deep Worker

• Purpose: Goal-oriented autonomous problem-solving
• Key Features:
  • Explores before acting (2-5 parallel explore/librarian agents)
  • End-to-end completion with verification
  • Pattern matching for codebase consistency
  • Surgical, minimal code generation
  • Thorough research before action

Oracle — Architecture & Debugging

• Purpose: High-IQ strategic consultation
• Key Features:
  • Complex debugging after 2+ failed attempts
  • Architecture design decisions
  • Multi-system tradeoffs analysis
  • Security/performance concerns review
  • Read-only consultation model

Librarian — Documentation & Search

• Purpose: External knowledge retrieval
• Key Features:
  • Official documentation search
  • Open source implementation patterns
  • Real-time source code digestion
  • Best practices for unfamiliar libraries
  • Multi-source search (docs, GitHub, Stack Overflow)

Explore — Fast Codebase Exploration

• Purpose: Rapid codebase analysis
• Key Features:
  • Blazing fast contextual grep
  • Pattern discovery across modules
  • Multiple search angles simultaneously
  • Quick file and symbol location

2. Multi-Model Orchestration

Model Selection

• Support for multiple LLM providers:
  • OpenAI (GPT-4o, GPT-4 Turbo, etc.)
  • Anthropic (Claude 3.5, Claude 3, etc.)
  • Google (Gemini Pro, etc.)
  • OpenCode Zen (curated models)

Fallback Chains

• Automatic fallback on model failure
• Configurable fallback order
• Per-agent model selection

Model Optimization

• Temperature tuning per agent
• Context window optimization
• Token usage tracking

3. Background Task Execution

Parallel Execution

• Run multiple agents simultaneously
• Async task completion
• Concurrent tool execution
• Background task persistence

Task Management

• Task queue with prioritization
• Progress tracking and reporting
• Task cancellation and restart
• Background/foreground switching

4. LSP & AST Tools

Built-in LSP Tools

• lsp_diagnostics: Get errors, warnings, hints
• lsp_goto_definition: Jump to symbol definition
• lsp_find_references: Find all usages of symbol
• lsp_rename: Rename symbol across workspace
• lsp_symbols: Get symbols from file/workspace
• lsp_prepare_rename: Check if rename is valid

AST-Grep Tools

• ast_grep_search: AST-aware code search
• ast_grep_replace: AST-aware code replacement
• Multi-language support (25+ languages)

Refactoring Capabilities

• Deterministic, surgical refactoring
• Safe multi-file changes
• Pattern-based code transformation

5. Hash-Anchored Edit Tool

How It Works

• Uses LINE#ID format for edits
• Validates content hash before applying changes
• Prevents stale-line edit errors
• Atomic, verifiable modifications

Usage Pattern

// Before: Line 42 contains specific content
// Edit is anchored to content hash, not line number
// Ensures edit applies to intended location

6. MCP (Model Context Protocol) Integration

Built-in MCPs

Custom MCPs

• Add custom MCP servers
• Configure MCP groups
• Per-project MCP settings
• MCP server authentication

7. Hook System

Available Hooks

Hook Configuration

• Enable/disable specific hooks
• Chain multiple handlers
• Custom hook scripts
• Per-agent hook settings

8. Claude Code Compatibility

Supported Features

• Commands: / commands for common actions
• Agents: Full agent system compatibility
• Skills: Skill tool integration
• MCP: MCP server integration
• Hooks: PreToolUse, PostToolUse, UserPromptSubmit, Stop

Compatibility Mode

• Seamless migration from Claude Code
• Identical command interface
• Shared configuration patterns

9. Session Management

Session Features

• Session List: View all sessions
• Session Read: Review previous sessions
• Session Search: Find by content
• Session Export: Share or backup

Session Persistence

• Configurable storage location
• Automatic session saving
• Session continuity across restarts
• Max session count management

10. Todo System

Features

• TODO List: Track all tasks
• TODO Write: Create/manage todos
• Status Tracking: Pending, in_progress, completed
• Auto-Continuation: Forces agent to finish

Enforcement

• Strict TODO following
• Auto-reminder for incomplete tasks
• Task completion verification
• Progress reporting

Productivity Features

• Core operation execution with comprehensive error handling
• Input validation and output quality assurance
• Integration with existing workflows and toolchains
• Detailed logging for debugging and audit trails

1. Ralph Loop

Purpose: Self-referential development loop

Features:

• Automatic plan execution and validation
• Iterative improvement
• Goal tracking
• Progress monitoring

2. Ultrawork Mode

Activation: Include "ultrawork" or "ulw" in prompt

Effects:

• Parallel agents activate
• Background tasks enabled
• Deep exploration mode
• Relentless execution

3. Comment Checker

Purpose: Maintain code quality

Rules:

• Prevents excessive comments
• Justifies necessary comments
• Enforces code readability
• Follows project style

4. Context Injection

Features:

• Auto-inject AGENTS.md
• Conditional rules injection
• Project context awareness
• User preferences preservation

5. Auto-Completion

Features:

• End-to-end task completion
• Verification before finishing
• Evidence-based reporting
• Quality gate enforcement

Workflow Enhancements

1. Receive input and validate format
2. Route to appropriate handler based on input type
3. Execute core operation with monitoring
4. Transform output to expected format
5. Return results or trigger follow-up actions

1. Parallel Agent Dispatching

// Fire multiple agents for independent tasks
- Explore: Map codebase structure
- Librarian: Find relevant documentation
- Hephaestus: Research best practices
- Oracle: Review architecture

2. Subagent-Driven Development

// Dispatch fresh subagent per task
- Two-stage review (spec, quality)
- No context pollution
- Parallel-safe execution
- Review checkpoints

3. Git Worktree Integration

// Isolated workspace per task
- Create worktree for feature
- Commit to feature branch
- Merge when complete
- Clean up when done

4. Test-Driven Development

// Subagents follow TDD
- Write test first
- Implement feature
- Verify tests pass
- Refactor as needed

5. Systematic Debugging

// Structured debugging approach
- Reproduce issue
- Identify root cause
- Implement fix
- Verify solution
- Document learnings

Tool Categories

File Operations

• read: Read file contents
• write: Write/overwrite files
• edit: Modify file contents
• glob: Find files by pattern
• grep: Search file contents

Code Analysis

• lsp_diagnostics: Get diagnostics
• lsp_goto_definition: Jump to definition
• lsp_find_references: Find references
• lsp_rename: Rename symbol
• ast_grep_search: AST search
• ast_grep_replace: AST replace

System Operations

• bash: Execute shell commands
• interactive_bash: Interactive terminal
• session_*: Session management
• Notion_*: Notion integration

Browser Operations

• skill_mcp: MCP server tools
• playwright: Browser automation
• dev-browser: Persistent browser state
• agent-browser: CLI browser automation
• browser-use: Web automation

Web Operations

• websearch_web_search_exa: Web search
• google_search: Google search + analysis
• webfetch: Fetch URL content
• look_at: Analyze media files

Configuration Categories

{
  "enabled": true,
  "autoRun": false,
  "timeout": 30000,
  "retries": 3
}

Set environment variables as needed for authentication and endpoints.

Agent Configuration

• Model selection per agent
• Temperature tuning
• Permission settings
• Custom prompts

Feature Configuration

• Parallel execution limits
• Background task settings
• Hook enable/disable
• Quality gate settings

Integration Configuration

• MCP server setup
• Hook script configuration
• Plugin integration
• External tool connection

UI Configuration

• Theme selection
• Keybind customization
• Layout preferences
• Display options

Security Features

• Core operation execution with comprehensive error handling
• Input validation and output quality assurance
• Integration with existing workflows and toolchains
• Detailed logging for debugging and audit trails

Permission System

• File system access control
• Network access restrictions
• Shell command allowlisting
• API key management

Validation

• Input validation before execution
• Content hash verification
• Safe mode options
• Audit logging

Compliance

• ToS-aware configuration
• Security best practices
• Audit trail maintenance
• Secret management

Performance Features

• Core operation execution with comprehensive error handling
• Input validation and output quality assurance
• Integration with existing workflows and toolchains
• Detailed logging for debugging and audit trails

Optimization

• Intelligent context truncation
• Token usage tracking
• Parallel execution efficiency
• Cache management

Scalability

• Concurrent agent limits
• Resource usage monitoring
• Memory optimization
• Background task queuing

Extensibility

This section covers extensibility for the oh-my-opencode-features skill. Key operations include input validation, core processing, and output verification. Refer to the skill overview for detailed usage instructions.

Custom Tools

• Create custom tools
• Define tool schemas
• Implement execution logic
• Register with system

Custom Hooks

• Write custom hooks
• Script integration
• Event-driven automation
• Workflow customization

Plugin System

• Extend functionality
• Add new agents
• Integrate external services
• Customize behavior

Comparison with Other Harnesses

Feature Quick Reference

The Process

1. Identify feature needs – Determine which oh-my-opencode features address your task (agents, MCPs, hooks, sessions, workflows)
2. Verify feature availability – Check feature is enabled in configuration and compatible with your OpenCode version
3. Configure feature – Set up MCP servers, hooks, categories as needed in config
4. Load feature in session – Ensure feature is accessible (agents via /agent, MCPs via /mcp, etc.)
5. Use feature – Execute feature-specific workflows (parallel agents, hooks, ultrawork, etc.)
6. Monitor execution – Watch feature behavior, adjust configs, debug issues
7. Verify output – Run lsp_diagnostics, build, tests to confirm feature completed correctly

When NOT to Use

• Task is outside your authorization scope
• You need to implement controls (use implementing-* skills)
• Task is about analysis, not action (use analyzing-* skills)
• You don't have access to target systems
• Task requires compliance expertise (consult professionals)
• Task is about defense, not offense (use defensive skills)

Red Flags

• Feature conflicts – Enabling multiple conflicting features (e.g., multiple LSP tools fighting over same file)
• Version mismatch – Using features from different oh-my-opencode versions causing incompatibilities
• Missing dependencies – Enabling MCPs without required tools or permissions
• Hook loops – Hooks triggering themselves infinitely due to improper conditions
• Over-enforcement – Enabling all quality checks when only specific ones needed causes slowdowns
• Ignoring workflows – Not following Ultrawork, Ralph Loop, or TDD when they should apply
• Session corruption – Attempting to use features on stale or corrupted sessions

Verification

• Feature loads: Feature commands work (e.g., /mcp list, /agent sisyphus)
• MCPs respond: MCP servers execute tool calls and return expected results
• Hooks fire: Hook conditions trigger expected actions at right times
• Sessions valid: Session commands return valid data without errors
• Workflows execute: Ultrawork completes autonomously, TDD creates tests, Ralph Loop iterates to completion
• Quality checks pass: Comment checker, TODO enforcement, verification run as configured
• Integration clean: All features work together without conflicts or race conditions

References

• Features Overview: https://github.com/code-yeongyu/oh-my-opencode#features
• Configuration: https://github.com/code-yeongyu/oh-my-opencode/blob/master/docs/configurations.md
• Agent System: https://github.com/code-yeongyu/oh-my-opencode#for-those-who-want-to-read-meet-sisyphus
• OpenCode Docs: https://opencode.ai/docs/

Anti-Rationalization Table

Related Skills

• oh-my-opencode: Overall integration
• oh-my-opencode-installation: Setup guide
• oh-my-opencode-agents: Agent deep dive
• oh-my-opencode-usage: Daily usage
• oh-my-opencode-configuration: Advanced setup

Process

1. Analyze the task requirements
2. Apply domain expertise
3. Verify output quality