Codex GodMode ON — AI skill for Claude Code
GPT-5.6-ready local Codex orchestration with 14 agents, 10 core skills, an optional local-first Paperwork plugin, safe installers, bounded subagents, and explicit quality gates.
How to install Codex GodMode ON
This entry records only its repository, not the path inside it, so there is no
exact command to give. Open cubetribe/CODEX_GodMode_ON and copy the folder into
~/.claude/skills/, or the file into ~/.claude/agents/.
What Codex GodMode ON does
GPT-5.6-ready local Codex orchestration with 14 agents, 10 core skills, an optional local-first Paperwork plugin, safe installers, bounded subagents, and explicit quality gates.
Alternatives in AI
- Develop — Development phase - Build solutions with multi-AI implementation and quality gates 2.8k ★
- Loushang — AI-native agent harness for coding workflows by python: multi-model LLM orchestration, stateful sessions, tool 1.3k ★
- Pilotfish — Multi-model orchestration layer for Claude Code — the frontier model plans, cheaper models execute, verificati 688 ★
README
Codex GodMode
GodMode is a lean, globally installable orchestration layer for Codex. It keeps authority, worktree safety, one-writer delivery, recoverable migration, and observable validation explicit without prescribing every reasoning step.
The current core release is **GodMode `3.0.0`**. It is a major rebuild for strong parent models such as GPT-5.6 Sol: less standing context, fewer generic roles, deliberate specialist selection, and more deterministic verification. Read the comprehensive [GodMode 3.0 release document](docs/releases/3.0.0.md) for the rationale, breaking changes, measured package reductions, migration contract, rollback, and release evidence.
What changed in GodMode 3
- 7 optional custom agents instead of 14
- 9 independently triggered core skills instead of companion chains
- zero subagents by default, at most two selected specialists, and a packaged base-config thread cap of two
- parent or one built-in worker as the only tracked-file writer
- separate static
validatorand executabletesterresponsibilities - risk-based gates instead of mandatory double validation
- safe retirement of exact GodMode 2.0 assets with verified backups
- separate package checks, installer tests, capability diagnostics, and routing evals
The most important behavioral change is the default: the parent works alone. It delegates only a bounded, independent uncertainty that materially improves the result, and uses at most two specialists. A packaged specialist is selected with its registered `agent_type`; a matching task name alone does not activate the role.
Why the system became smaller
GodMode 2.x protected useful invariants, but repeated them through 14 custom agents, ten skills, a large workflow prompt, mandatory double validation, and a post-gate Scribe. With a stronger orchestrator, those layers could duplicate discovery, planning, implementation, and review while filling context and multiplying child turns.
GodMode 3 keeps the rules that a
Related Skills
Vibe Coding Toolkit
A curated, battle-tested AI-coding toolkit: Claude Code plugins, subagent orchestration, quality gates, and re
Dsh Legion
Multi-agent orchestration and LLM model routing for DeepSeek Harness (DSH): semantic AI agent profiles, exact
Ataru
High-performance AI memory retrieval for local agent history — a Rust search core (SDK / API / JSON CLI) plus
CodeEraser
An eraser against LLM-induced code & document entropy — clone/duplication/deadcode/structure judgment (Haskell
Tausik Core
AI coding agents can't quietly fake "done". Signed ed25519 receipts prove the gates actually ran — bound to th
Salient Core
Guardrail and permission layer for AI agent harnesses. A default-deny control kernel that gates every tool cal
Related Agents
Codex Orchestrator
CLI tool for delegating tasks to GPT Codex agents via tmux sessions. Designed for Claude Code orchestration wi
Codex Debugger
Root-cause a failing test, crash, stack trace, or misbehaving feature by delegating the investigation to OpenA
CI Local Runner
Use this agent to run a repo's CI gates locally and report whether the code is safe to push — especially when