Orchestrate multiple Claude Code workers with persistent state and governance protocols.
npx skills add https://github.com/cj-vana/claude-swarm --skill swarm-cj-vana
Please help me install this Agent Skill. Skill: swarm Source: https://github.com/cj-vana/claude-swarm/tree/main/skill Command: npx skills add https://github.com/cj-vana/claude-swarm --skill swarm-cj-vana
This Skill centralizes the orchestration of long-running coding sessions, enabling teams to run multiple Claude Code workers under governance with persistent state across context changes.
To begin, initialize a swarm session with a task description, optionally register governance protocols, then start workers and monitor progress. When features complete, run verifications, commit progress, and review results.
High-intent search queries and answers about installing and using this skill.
You can orchestrate multiple Claude Code workers by initializing a swarm session with a task description and governance protocols. This enables parallel workers to accelerate complex refactors and multi-step initiatives while enforcing behavioral constraints.
Protocol-governed coding session orchestration is the centralization of long-running coding tasks under behavioral constraints. It applies safety and compliance rules to multiple workers, ensuring structured completion, review, and rollback workflows for large feature delivery.
Yes, you can preserve session progress across context compactions. The orchestration skill maintains persistent state for decisions and progress, ensuring that long-running coding sessions survive context changes and restarts without losing data.
The best way to manage parallel coding sessions with persistent state is using an orchestration skill that registers governance protocols and monitors workers. This approach supports structured reviews and rollback workflows for complex refactors.
To start a swarm session, initialize it with a task description and optionally register governance protocols. Then start workers, monitor progress, run verifications when features complete, commit progress, and review results.
You should avoid using parallel workers for multi-step coding tasks if your project lacks clear behavioral constraints or governance protocols. Without structured completion and rollback workflows, orchestrating concurrent sessions may lead to non-compliant or unsafe code changes.
cristoslcOrchestrate Claude Code workers in tmux panes for parallel task execution.
oneforceExecute PRD-driven multi-iteration workflows with reviewer verification and progress tracking.
Rune-kitMonitor tool call counts and trigger state saving before context compaction.
zhixuan312Dispatch Markdown plan tasks sequentially to a single worker session.
