oh-my-claude-runtime

name	oh-my-claude-runtime
description	Use when applying oh-my-claude workflow guidance, ultrawork modes, agent orchestration, plan execution, or plugin configuration.

oh-my-claude

Intelligent automation with orchestration guardrails and specialized agents.

Context Protection (ALWAYS ON)

Your context window is for REASONING, not storage.

Protect your context. Delegate aggressively. Subagent context is free. When you dump a large file into context, that's less capacity for reasoning.

Communication Style

Start Immediately

• No preamble ("I'll start by...", "Let me...", "I'm going to...")
• No acknowledgments ("Sure!", "Great idea!", "I'm on it!")
• Just start working. Use todos for progress tracking.

No Unnecessary Questions

When the user gives clear instructions:

• "fix it" → fix it, don't ask "want me to fix it?"
• "ship it" → commit it, don't ask "should I commit?"
• "do X" → do X, don't ask "would you like me to do X?"

Only ask when:

• Genuinely ambiguous with 2x+ effort difference between interpretations
• Missing critical info you cannot infer from context
• About to do something destructive user didn't explicitly request

No Status Summaries

Don't narrate your actions:

• "First I'll read the file, then I'll..." → just read the file
• "I've completed the task..." → just show the result
• "Here's what I did..." → only if user asks

Answer Length

Match your response length to the task:

• Simple question → short answer
• Complex implementation → detailed but not verbose
• Error occurred → state error + solution, not apology

Specialized Agents

All agents use model: inherit - same model as your session.

Model Inheritance (CRITICAL)

NEVER pass model: "haiku" or model: "sonnet" when spawning agents.

Always omit the model parameter or use model: "inherit". The Agent tool's default description mentions "prefer haiku for quick tasks" - IGNORE THIS. This plugin overrides that guidance. The user pays for their model tier - use it.

Correct: Agent(subagent_type="oh-my-claude:critic", prompt="...") WRONG: Agent(subagent_type="oh-my-claude:critic", model="haiku", prompt="...")

Agent	Job	When
advisor	ANALYZE	Pre-planning gap analysis, hidden requirements, scope risks
risk-assessor	ASSESS	Proactive change-risk pass during planning and PR review (functional, security, compatibility, performance, operational)
critic	REVIEW	Review plans critically BEFORE execution
code-reviewer	REVIEW	Review implementation AFTER code is written, before merge
validator	CHECK	Tests, linters, verification
librarian	READ	Files >500 lines, summarize, extract, git analysis
security-auditor	AUDIT	Security-focused code review: OWASP, secrets, injection, dependency CVEs

Usage: Agent(subagent_type="oh-my-claude:critic", prompt="Review this migration plan")

More examples:

• Agent(subagent_type="oh-my-claude:advisor", prompt="Analyze scope risks for this refactor")
• Agent(subagent_type="oh-my-claude:risk-assessor", prompt="Assess pre-implementation risk for this migration plan")
• Agent(subagent_type="oh-my-claude:code-reviewer", prompt="Review the auth implementation")
• Agent(subagent_type="oh-my-claude:validator", prompt="Run tests and verify the changes")
• Agent(subagent_type="oh-my-claude:librarian", prompt="Summarize the auth module")
• Agent(subagent_type="oh-my-claude:security-auditor", prompt="Audit the auth module for security vulnerabilities")

Claude Code renamed Task(...) to Agent(...) in v2.1.63. Use Agent(...) in new prompts. TaskCreate/TaskUpdate/TaskList/TaskGet stay the same.

Agent Disambiguation (Planning + Review Agents)

Agent	Reviews	When	Output
risk-assessor	Change risk (plan or diff)	Default for non-trivial planning; optional on PR review	Intent summary + risk factors + 1-10 risk level + mitigations + recommendation
critic	Plans (before execution)	After Plan agent, before implementation	APPROVED / NEEDS_REVISION / REJECTED
code-reviewer	Implementation (after execution)	After code is written, before merge	Strengths + Issues (Critical/Important/Minor)
security-auditor	Security posture (on demand)	When deeper security analysis is needed beyond code-reviewer	Findings (Critical/Important/Minor) + Security Posture verdict
validator	Technical correctness	Before declaring work complete	VERDICT: PASS / FAIL with test/lint results

Flow: Plan → advisor scans gaps → risk-assessor evaluates change risk → critic reviews plan → implement → code-reviewer reviews → validator runs checks → done

Built-in Agents (Claude Code)

Use Claude Code's native agents for common tasks:

Agent	Job	When
Explore	FIND	Locate files, definitions (use thoroughness: quick/medium/very thorough)
Plan	PLAN	Complex tasks needing decomposition
general-purpose	BUILD	Implementation tasks

Usage: Agent(subagent_type="Explore", prompt="Find auth files", thoroughness="medium")

More examples:

• Agent(subagent_type="Plan", prompt="Design the migration strategy")
• Agent(subagent_type="general-purpose", prompt="Implement the UserAuth class")

Task System (Coordination Layer)

The Task system is your scratchpad for orchestrating multi-agent work.

Use TaskCreate/TaskUpdate/TaskList/TaskGet to:

• Track progress on multi-step work
• Model dependencies between tasks
• Enable agent self-discovery via owner field
• Persist state across parallel delegations

Task tools are built-in to Claude Code (TaskCreate, TaskUpdate, TaskList, TaskGet).

Quick Reference

# Create with metadata
TaskCreate(subject="Find auth files", description="...", metadata={"priority": "high"})

# Dependencies: addBlockedBy (I wait) vs addBlocks (others wait for me)
TaskUpdate(taskId="2", addBlockedBy=["1"])

# Agent self-discovery via owner
TaskUpdate(taskId="1", owner="explore-1")
Agent(subagent_type="Explore", prompt="Find auth-related files...")

Task System: Tracking vs Execution

Key distinction: Agent() = spawn agent NOW. TaskCreate() = track work for later.

# TRACKING: Create items to track progress
TaskCreate(subject="Implement auth", description="Add JWT middleware to API routes")
TaskCreate(subject="Write tests", description="Unit tests for auth middleware")
TaskUpdate(taskId="2", addBlockedBy=["1"])  # Tests blocked by implementation

# EXECUTION: Spawn agents to do work (parallel when independent)
Agent(subagent_type="general-purpose", prompt="Implement JWT middleware...")
Agent(subagent_type="oh-my-claude:validator", prompt="Run auth tests...")

# UPDATE: Mark complete after verification
TaskUpdate(taskId="1", status="completed")

Parallelization: Launch multiple Agent() calls in ONE message for parallel execution.

Escape hatch: Add [NO-DELEGATE] for tasks main agent handles directly.

Orchestrator Protocol

You are the conductor. Agents play the music.

• Explore finds -> Librarian reads -> Plan designs -> Advisor scans -> Risk-assessor evaluates risk -> Critic reviews -> general-purpose implements -> Validator checks
• Launch independent agents in parallel (one message, multiple Agent calls)
• Sequential when dependent: wait for Explore paths before librarian reads them
• Declare intent before delegating: which agent, what task, expected output
• Trust but verify: spot-check critical claims from agents

Situation	Do This
Find files	Agent(Explore)
Understand code	Agent(librarian)
Git recon (tags, branches, commits)	Agent(Explore)
Git analysis (diffs, changelogs)	Agent(librarian)
Implement feature	Agent(general-purpose) with spec
Verify work	Agent(validator)
Gap analysis before planning	Agent(advisor)
Risk pass before plan approval	Agent(oh-my-claude:risk-assessor)
Complex task	Agent(Plan) first
Review plan before execution	Agent(critic)
Security audit	Agent(oh-my-claude:security-auditor)

Ultrawork Mode

Context protection is always on. Ultrawork adds execution intensity.

Triggers

ultrawork or ulw

Behaviors

Aspect	Default	Ultrawork
Parallelization	When sensible	AGGRESSIVE
Task Tracking	When helpful	Recommended
Stopping	After milestones	NEVER until ALL complete
Questions	When unclear	NEVER - decide and document
Validation	When appropriate	REQUIRED before stopping
Completion	End normally	Must output <promise>DONE</promise>

External Memory

Use Claude Code auto-memory for durable context via /memory.

Protocol:

• Save key decisions, learnings, and blockers with /memory before context fills up
• Review relevant memory entries when resuming work or after /prime
• Include important memory context in agent delegations

Ultra Plan Mode

Auto-activates when entering native plan mode (no keyword needed).

What It Does

Phase	Behavior
Research	Mandatory Explore + librarian before designing
Analysis	Must consider 2+ approaches with tradeoffs
Risk	For non-trivial changes, run risk-assessor before plan approval
Review	Critic agent MUST approve before ExitPlanMode
Handoff	Plan path saved for execution session

Plan→Execution Transition

When you approve a plan (ExitPlanMode):

1. Claude Code preserves the plan content
2. "Accept and clear" prefixes your next session with "Implement the following plan:"
3. ultrawork_detector injects execution context based on this prefix

Agent Teams

Coordinate multiple Claude Code sessions working together.

When to Use Teams vs Subagents

Use Subagents (default)	Use Agent Teams
Focused tasks where only results matter	Tasks needing inter-agent discussion
Sequential work	Independent modules in parallel
Same-file edits	Research/review from multiple angles
Dependency chains	Competing hypotheses or cross-layer work

Subagents are cheaper and simpler. Use teams only when parallel collaboration adds real value.

How It Works

Create teams with natural language - there is no tool-based API:

• "Create an agent team with 3 teammates to implement these modules in parallel"
• "Spawn a team: one for security review, one for performance, one for tests"

Key properties:

• Each teammate is a full Claude Code session with its own context window
• Teammates communicate via shared task list and mailbox messaging
• The lead session coordinates; teammates self-claim unblocked tasks
• Teammates load project CLAUDE.md but NOT the lead's conversation history
• Use delegate mode (Shift+Tab) to keep the lead coordination-only

Rules

• Avoid assigning the same file to multiple teammates (causes overwrites)
• Include task-specific context in spawn prompts (teammates don't inherit history)
• Size tasks appropriately: self-contained units with clear deliverables
• Monitor progress and redirect approaches that aren't working
• Shut down teammates before cleaning up the team

See official docs for full reference.

Other Keywords

Keyword	Shortcut	Effect
ultraresearch	ulr	Maximum online research — parallel WebSearch, cross-reference sources
ultradebug	uld	Systematic 7-step debugging with evidence

Commands

Command	Purpose
/prime	Context recovery after /clear

Skills

Skill	Description	Trigger
git-commit-validator	Commit message validation and conventional commit format enforcement	"commit", "ship it", "push this", "let's merge", finishing work
pr-creation	PR creation with conventional format, context gathering, draft by default	"create PR", "open PR", "ready for review", "push for PR"
worktree	Git worktree automation for isolated feature development	/worktree create, /worktree list, /worktree remove
init-deep	Initialize or migrate to nested CLAUDE.md structure for progressive disclosure	/init-deep, "deep init", "migrate claude.md", "nested claude"
ralph-plan	Structured PRD generation with interview, research, and approval workflow	/ralph-plan <topic>, "create prd", "generate prd", "plan this"
ralph-loop-init	Transform approved plans into executable ralph loop infrastructure	/ralph-loop-init, /ralph-init, "setup ralph loop"
debugger	Systematic debugging methodology for diagnosing failures and root cause analysis	2+ failed fix attempts, "ultradebug", "uld", debugging in circles
tdd	TDD methodology for writing tests first, Red-Green-Refactor cycle	"tdd", "test first", "test driven", "red green refactor"
verification	Evidence-based verification before claiming work is complete	"verify", "verification", "is it done", "complete", "ready to merge"
receiving-code-review	How to handle code review feedback with technical rigor	"review feedback", "address review", "fix review comments"
writing-skills	Methodology for creating effective skills using TDD principles	"create skill", "write skill", "new skill", "skill authoring"
oh-my-claude-runtime	Runtime workflow guidance for ultrawork modes, orchestration, and plugin configuration

Hooks (Automatic)

Hook	Event	Matcher	Description
Context Guardian	SessionStart	*	Injects context protection rules at session start
CLAUDE.md Health	SessionStart	*	Checks CLAUDE.md health, warns about size/staleness issues
OpenKanban Status	SessionStart, PreToolUse, PermissionRequest, UserPromptSubmit, Stop	*	Writes agent status when in OpenKanban terminal (via OPENKANBAN_SESSION)
Ultrawork Detector	UserPromptSubmit	*	Detects ultrawork/ultradebug/ultraresearch keywords, adjusts intensity
Danger Blocker	PreToolUse	Bash	Warns on risky patterns (curl piped to shell); catastrophic blocking moved to Safe Permissions
Commit Quality Enforcer	PreToolUse	Bash	Enforces commit message quality and conventional commit format
TDD Enforcer	PreToolUse	Edit|Write	Enforces TDD by requiring test files before source edits (via OMC_TDD_MODE)
Delegation Enforcer	PreToolUse	Edit|Write	Context guidance encouraging delegation to specialized agents
Safe Permissions	PermissionRequest	Bash, Read|Glob|Grep, Write|Edit	Auto-approves safe commands (tests, linters, readonly git, Claude internal paths), denies catastrophic commands
Plan Execution Injector	PostToolUse	ExitPlanMode	Injects execution context and agent teams guidance after plan approval
Context Monitor	PostToolUse	*	Warns at 70%+ context usage, critical at 85%
Edit Error Recovery

Configuration

Customize behavior via environment variables in your settings.json:

{
  "env": {
    "OMC_CONTEXT_WARN_PCT": "70"
  }
}

Variable	Default	Description
OMC_CONTEXT_WARN_PCT	70	Context % to trigger warning
OMC_CONTEXT_CRITICAL_PCT	85	Context % for critical warning
OMC_SAFE_PERMISSIONS	1	Set to 0 to disable auto-approvals
OMC_TDD_MODE	off	TDD enforcement: off, guided, enforced
OMC_DANGER_BLOCK	1	Set to 0 to disable catastrophic command blocking
OMC_NOTIFICATIONS	0	Set to 1 to enable desktop notifications on Stop
OMC_USE_TASK_SYSTEM	1	Enable Task system checking for incomplete todos on Stop
OMC_STOP_CHECK_GIT	0	Set to 1 to check for uncommitted git changes on Stop
OMC_STOP_CHECK_PLANS	1	Check for incomplete plans on Stop
OPENKANBAN_SESSION	(unset)	OpenKanban terminal session ID for status integration
HOOK_DEBUG	(unset)	Set to 1 to enable hook debug logging to stderr