Multi-agent orchestration for Claude Code, ported from oh-my-openagent. Greek mythology-themed agents for planning, execution, research, and review.
Plan executor and orchestrator that delegates ALL implementation via sub-agents. Never writes code itself. Reads plans, builds parallelization maps, delegates tasks, and rigorously verifies every result. (Atlas - SisyClaude Plugin)
Fast codebase search specialist. Answers "Where is X?", "Which file has Y?", "Find the code that does Z". Fire multiple in parallel for broad searches. (Explore - SisyClaude Plugin)
Autonomous deep worker that never asks permission and keeps going until 100% done. Senior Staff Engineer persona. Use for deep implementation tasks, complex bug fixes, and end-to-end feature work. Fires explore/librarian agents in parallel for research. (Hephaestus - SisyClaude Plugin)
Open-source codebase understanding agent for documentation lookup, library research, and finding implementation examples using GitHub CLI, web search, and documentation fetching. (Librarian - SisyClaude Plugin)
Pre-planning consultant that analyzes requests to identify hidden intentions, ambiguities, and AI failure points before planning begins. Use for complex or ambiguous requests. (Metis - SisyClaude Plugin)
Activate SisyClaude by writing its system prompt to ~/.claude/SISYCLAUDE_SYSTEM_PROMPT.md and adding a `sisyclaude` shell alias that loads it via `claude --system-prompt-file` (replaces the default Claude Code system prompt for that session).
Challenge your design ideas before committing to a plan. Spawns the Socrates agent for adversarial-constructive dialogue that stress-tests assumptions, surfaces trade-offs, and sharpens your thinking.
Deactivate SisyClaude by removing the `sisyclaude` shell alias and deleting ~/.claude/SISYCLAUDE_SYSTEM_PROMPT.md.
Uses power tools
Uses Bash, Write, or Edit tools
Own this plugin?
Verify ownership to unlock analytics, metadata editing, and a verified badge. GitHub access is read-only (username + org membership).
Sign in to claimOwn this plugin?
Verify ownership to unlock analytics, metadata editing, and a verified badge. GitHub access is read-only (username + org membership).
Sign in to claimBased on adoption, maintenance, documentation, and repository signals. Not a security audit or endorsement.
A port of oh-my-openagent's Greek mythology-themed agent system, brought back to Claude Code.
This project exists because of work that happened outside the Anthropic ecosystem — and the story is worth telling.
In early 2026, oh-my-openagent (originally "oh-my-opencode") by @code-yeongyu introduced a multi-agent orchestration layer for OpenCode — a model-agnostic open-source coding agent. The system used Greek mythology-themed agents (Prometheus for planning, Hephaestus for deep implementation, Oracle for architecture decisions, etc.) to route tasks to the right specialist. It grew to 43k+ GitHub stars and became one of the most popular plugins in the OpenCode ecosystem.
Around the same time, Anthropic moved to block third-party tools from using Claude subscription credentials via OAuth, citing terms-of-service violations around identity spoofing and token routing. This action — which broke OpenCode, Cline, RooCode, and others overnight — was partly attributed to oh-my-openagent's success. The dispute escalated through formal legal demands, and by March 2026 OpenCode had removed all Anthropic integration entirely.
The oh-my-openagent community built something genuinely innovative during and despite that conflict. Their multi-agent architecture — the idea that a coding assistant should delegate rather than do everything in one context window — proved itself across thousands of real-world projects.
SisyClaude takes the agent architecture and design philosophy pioneered by oh-my-openagent and adapts it for Claude Code's native sub-agent system. Instead of TypeScript orchestration running outside Claude, the agents are defined as markdown agent definitions that Claude Code loads natively.
The name "SisyClaude" is intentional — this is a reduced, Claude-specific adaptation, not a replacement for the original. Think of it as a tribute, not a competitor.
What works well:
What oh-my-openagent does better:
If you want the full experience with provider flexibility, go use oh-my-openagent. This port is for people who prefer (or need) to stay within Claude Code and want better orchestration than the vanilla experience.
Sisyphus — Main orchestrator: classifies intent, delegates to specialists, verifies results. Runs at the top level (not a sub-agent, because the Agent tool cannot nest) and is loaded into Claude Code via the sisyclaude shell alias added by /sisyclaude:activate. Run sisyclaude instead of claude to start a session with Sisyphus's instructions in effect.
| Agent | Role | Model | Invocation |
|---|---|---|---|
| Atlas | Plan executor — delegates ALL implementation, never codes | Opus | @atlas |
| Hephaestus | Autonomous deep worker — never asks, just completes | Opus | @hephaestus |
| Prometheus | Strategic planner — interviews, researches, writes plans | Opus | @prometheus |
| Oracle | Architecture consultant — read-only, effort-tagged advice | Opus | @oracle |
| Momus | Plan reviewer — approval-biased, max 3 issues per rejection | Sonnet | @momus |
| Metis | Pre-planning analyst — catches hidden requirements, AI slop | Opus | @metis |
| Explore | Fast codebase search — structured results, parallel execution | Haiku | @explore |
| Librarian | OSS docs researcher — evidence-based with GitHub permalinks | Sonnet | @librarian |
| Socrates | Design critic — adversarial-constructive idea stress-testing | Opus | @socrates / /brainstorm |
| Skill | What it does |
|---|---|
/brainstorm | Spawns Socrates for interactive design challenge dialogue |
/sisyclaude:activate | Activates SisyClaude as default orchestrator |
/sisyclaude:deactivate | Deactivates SisyClaude, restores original CLAUDE.md |
/plugin marketplace add Whiskey-Tango-Foxtrot-GmbH/sisyclaude
/plugin install sisyclaude@sisyclaude
/sisyclaude:activate
npx claudepluginhub whiskey-tango-foxtrot-gmbh/sisyclaude --plugin sisyclaudeThis skill should be used when the model's ROLE_TYPE is orchestrator and needs to delegate tasks to specialist sub-agents. Provides scientific delegation framework ensuring world-building context (WHERE, WHAT, WHY) while preserving agent autonomy in implementation decisions (HOW). Use when planning task delegation, structuring sub-agent prompts, or coordinating multi-agent workflows.
Verification-first engineering toolkit for Claude Code. 15 skills across a 5-phase spine (Investigate → Design → Implement → Verify → Ship), 8 specialist agents, an interactive setup wizard. Every skill has rationalizations + evidence requirements. Built for senior ICs and tech leads.
HelloAGENTS — The orchestration kernel that makes any AI CLI smarter. Adds intelligent routing, unified QA gates, safety guards, and notifications.
OpenAgentsControl — multi-agent orchestration for Claude Code. Context-aware development with skills, subagents, parallel execution, and automated code review.
Describe your goal, approve the spec, then step away — Claude and Codex loop together until it's right.
Delegate plan execution to Codex CLI via ASP. Part of cc-multi-cli-plugin. Requires the `multi` plugin.