polyphony

Orchestrate parallel coding agents in isolated Docker containers.

Updated Jun 19, 2026

Search Tags:#docker #parallel development #multi-agent orchestration #task routing #git branches #container isolation #agent workflows

Ready to run

npx skills add https://github.com/lciacci/tessera --skill polyphony-lciacci

💬 Prompt for ChatGPT / Claude Chat▼

Please help me install this Agent Skill.
Skill: polyphony
Source: https://github.com/lciacci/tessera/tree/main/skills/polyphony
Command: npx skills add https://github.com/lciacci/tessera --skill polyphony-lciacci

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

Polyphony removes the coordination overhead and workspace conflicts involved in running multiple coding agents on parallel tasks.

Core Features & Use Cases

Container-Isolated Execution: Give each agent an independent Docker workspace, git clone, branch, credentials, and test environment.
Automated Orchestration: Discover, claim, route, provision, run, verify, retry, and land tasks through a managed lifecycle.
Intelligent Routing: Match tasks to agent profiles using configurable rules, fallback chains, and five-dimension complexity scoring.
Use Case: Run several feature or bug-fix tasks concurrently while preserving clean branches and isolating failures between agent sessions.

Quick Start

Use the polyphony skill to initialize the orchestration configuration and spawn a container-isolated agent team for the requested coding tasks.

Frequently Asked Questions about polyphony

High-intent search queries and answers about installing and using this skill.

FAQPage Schema

How do I run parallel coding agents without git branch conflicts?▼

To run parallel coding agents without git branch conflicts, use container-isolated execution. This approach provisions independent Docker workspaces, git clones, and branches for each agent, preventing cross-task interference during concurrent feature development or bug fixes.

Can I orchestrate multiple AI agents like Claude and Codex concurrently?▼

Yes, you can orchestrate multiple AI agents concurrently. The system matches tasks to agent profiles like Claude, Codex, or Kimi CLIs using configurable routing rules, fallback chains, and complexity scoring to manage the concurrent execution lifecycle.

Do I need Docker to isolate agent workflows for local task queues?▼

Yes, Docker or OrbStack is required to isolate agent workflows for local task queues. The environment depends on these container platforms to provision independent workspaces, manage credentials, and execute test environments safely for each agent.

What is the best way to route GitHub issues to different AI coding agents?▼

The best way to route GitHub issues to different AI coding agents is using intelligent task routing. It applies a five-dimension complexity scoring system with configurable rules and fallback chains to match issue tasks to the appropriate configured agent profile.

How does container isolation handle agent retry and cleanup for failed tasks?▼

Container isolation handles agent retry and cleanup through a managed lifecycle state system. It automatically verifies task results, applies retry handling for failed agent sessions, and executes cleanup procedures to isolate failures from other concurrent workspaces.