Help us improve
Share bugs, ideas, or general feedback.
Share bugs, ideas, or general feedback.
Share bugs, ideas, or general feedback.
By Crysple
Orchestrate complex development tasks with a multi-agent team that uses contract-gated verification loops, form-driven cooperation, and structured phase orchestration to prevent premature exits and ensure spec compliance.
npx claudepluginhub crysple/superteam --plugin superteamYou are the **Architect**, responsible for decomposing the approved spec into incremental, independently verifiable parts, authoring frozen contracts with executable verification scripts, and adapting the plan throughout execution. You are a **teammate** in a Claude Code team (running in your own tmux pane), NOT a subagent. You communicate with other teammates via `SendMessage`.
You are the **Curator**, responsible for synthesizing session knowledge and persisting it to the global knowledge store at `~/.superteam/`. You are a **teammate** in a Claude Code team (running in your own tmux pane), NOT a subagent. You communicate with other teammates via `SendMessage`.
You are the **Explorer**, a shared research teammate responsible for deep codebase investigation and knowledge accumulation. You are a **teammate** in a Claude Code team (running in your own tmux pane), NOT a subagent. You communicate with other teammates via `SendMessage`.
You are the **Manager**, a stateless monitoring agent responsible for detecting anomalies, driving the execution loop, and escalating when patterns indicate problems. You are a **teammate** in a Claude Code team (running in your own tmux pane), NOT a subagent. You communicate with other teammates via `SendMessage`.
Pipeline orchestration agent - drives phase transitions, manages state, handles message routing for GATE-CHALLENGE/inability/completion, coordinates spawn requests through TL, and manages error recovery and restart cycles.
Executes bash commands
Hook triggers when Bash tool is used
Modifies files
Hook triggers on file write and edit operations
Uses power tools
Share bugs, ideas, or general feedback.
Own this plugin?
Verify ownership to unlock analytics, metadata editing, and a verified badge.
Sign in to claimOwn this plugin?
Verify ownership to unlock analytics, metadata editing, and a verified badge.
Sign in to claimBased on adoption, maintenance, documentation, and repository signals. Not a security audit or endorsement.
Dynamically assemble expert agent teams for complex tasks using Claude Code's agent teams feature
Orchestrate multi-agent teams for parallel code review, hypothesis-driven debugging, and coordinated feature development using Claude Code's Agent Teams
Launch agent teams for any kind of work β code, writing, or general tasks
Multi-agent orchestration for Claude Code. 12 specialized agents working in parallel β planning, building, reviewing, debugging. Plus a Hub for always-alive multi-project sessions controllable from Telegram or Slack.
Multi-agent team orchestration for parallel task execution, research, and implementation
Multi-agent team orchestration for Claude Code. Set up parallel AI agent teams with file-based planning, progress tracking, and role-based collaboration.
Uses Bash, Write, or Edit tools
Uses Bash, Write, or Edit tools
Share bugs, ideas, or general feedback.
one /superteam command spawns a real engineering team that brainstorms the spec with you, locks executable acceptance gates, and grinds through your enterprise stack (Flyte, HDFS, k8s, internal mirrors, your MCP servers) β increment by increment, overnight, until every gate passes.
Not another agent in a loop
No fake "looks good"
No re-teaching your company every Monday
Features Β· Quick Start Β· How It Works Β· Run on a Remote VM Β· Related Projects Β· Global Wiki
Language / θ―θ¨: English Β· δΈζ
Five things a single agent in a loop cannot do.
"Keep asking until your goal and the spec are byte-identical."
A real PM-style intake grounded in your codebase β not a vibes-based "got it, building now." The PM surveys what you already have, asks targeted classifying questions about scope, edge cases, and integration points, and refuses to move on while ambiguity remains. No more "the LLM made an assumption I'd never make" at hour 14 of a run.
"You approve scripts, not adjectives. After that, 'done' is binary β and no agent can negotiate it."
Acceptance criteria compile to executable shell scripts β pytest, ssh edge-01 'hdfs dfs -test ...', flytectl ... | jq -e '.phase=="SUCCEEDED"'. Review once, before any code. After that, "done" is a non-zero exit code or it isn't done. Adversarial reviewers can be sweet-talked. Failing exit codes cannot.
"Hand it off at 6 PM. Wake up to a green PR."
Claude Code alone caps at ~20 min before drift. Superteam runs 20+ hours: every increment gets a fresh Generator/Evaluator pair, and the Evaluator can't read the Generator's reasoning β adversarial by construction. No accumulated context. No fake "looks good." Pair with tmux on a remote VM and the team runs while you sleep.
"Real Isolated Claude Code sessions in tmux. A file-based harness keeps them in sync. Watchdogs bring them back."
Every teammate is a full Claude Code session in its own tmux pane β not an in-process sub-agent, not a multi-agent prompt. Coordination lives on disk (state.json, events.jsonl, frozen contracts, hooks), so the harness self-heals:
state.json goes staleInspired by Anthropic's harness-design writeup.
"Stop teaching it about your company every Monday."